27.01.2015 Views

Measuring Local Government Expenditure Needs - Social

Measuring Local Government Expenditure Needs - Social

Measuring Local Government Expenditure Needs - Social

SHOW MORE
SHOW LESS
  • No tags were found...

Create successful ePaper yourself

Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.

<strong>Measuring</strong> <strong>Local</strong> <strong>Government</strong><br />

<strong>Expenditure</strong> <strong>Needs</strong><br />

The Copenhagen Workshop 2007<br />

Editors: Junghun Kim and Jorgen Lotz


<strong>Measuring</strong> <strong>Local</strong><br />

<strong>Government</strong><br />

<strong>Expenditure</strong> <strong>Needs</strong> -<br />

The Copenhagen<br />

Workshop 2007<br />

Editors: Junghun Kim and Jorgen Lotz<br />

Publishers: The Korea Institute of Public Finance and the Danish<br />

Ministry of <strong>Social</strong> Welfare.<br />

The Copenhagen Workshop 2007 on 13 th and 14 th September 2007


Published by: The Korea Institute of Public Finance<br />

and The Danish Ministry of <strong>Social</strong> Welfare<br />

Year of publication: 2008<br />

Printing:<br />

Silkeborg Bogtryk<br />

Photo:<br />

iStockphoto<br />

Number printed: 600<br />

ISBN:<br />

978-87-754-6428-9 (printed)<br />

978-87-754-6429-6 (on-line)<br />

Price:<br />

DKK 150,00 plus postage<br />

If you would like to buy this publication:<br />

In Denmark please contact:<br />

SCHULTZ DISTRIBUTION A/S<br />

Herstedvang 10 A<br />

DK-2620 Albertslund<br />

Denmark<br />

E-mail:<br />

schultz@schultz.dk<br />

Phone: (+45) 43 22 73 00<br />

Fax: (+45) 43 63 19 69<br />

In Korea please contact:<br />

Korea Institute of Public Finance<br />

(79-6 Garak-dong)<br />

28 Bangjugmalgil Songpa-Ku<br />

Seoul<br />

Korea<br />

E-mail:<br />

pub@kipf.re.kr<br />

Phone: 82-2-2186-2144<br />

Fax: 82-2-2186-2179<br />

2


i. Foreword ..................................................................6<br />

ii. Opening address..........................................................8<br />

iii. Introduction............................................................. 12<br />

0. The book.................................................................................................... 12<br />

1. Is equalisation of needs necessary ........................................................... 13<br />

2. How do we measure expenditure needs................................................... 16<br />

3. The relationship to regional policy issues ................................................. 24<br />

4. Stability of grants over the years............................................................... 28<br />

5. The organisation of the measurement of local expenditure....................... 30<br />

6. Are earmarked grants an alternative to needs equalisation...................... 32<br />

1. Fiscal need equalization: Is it worth doing Lessons from<br />

international practices.................................................... 35<br />

1.1. Introduction ............................................................................................ 35<br />

1.2. Bridging the fiscal divide through fiscal equalization transfers –<br />

conceptual considerations ....................................................................... 36<br />

1.3. Designing fiscal equalization transfers................................................... 38<br />

1.4. Practical difficulties in equalizing expenditure needs: a review of<br />

worldwide experiences............................................................................ 46<br />

1.5. A simpler approach to expenditure need equalization: use of outputbased<br />

transfers for merit public services ................................................. 51<br />

1.6.Conclusions regarding the practice of fiscal equalization....................... 59<br />

2. <strong>Expenditure</strong> needs equalisation - reasoning and organisation of<br />

work: the Danish case..................................................... 61<br />

2.1. Introduction - the case for expenditure needs equalisation .................... 62<br />

2.2. What is the Danish reasoning behind equalisation of expenditure needs<br />

................................................................................................................. 66<br />

2.3. The organisation of work ....................................................................... 75<br />

2.4. Final comments ...................................................................................... 82<br />

3. The Use of <strong>Expenditure</strong> <strong>Needs</strong>: Equalization or Regional<br />

Policies ...................................................................... 87<br />

3.1. Introduction ............................................................................................ 87<br />

3. 2. Overview of local public finance .......................................................... 88<br />

3


3. 3. Structure of equalisation Grants ............................................................ 92<br />

3. 4. Discussions............................................................................................ 97<br />

3. 5. Conclusion........................................................................................... 104<br />

4. The equalised allocation of local expenditure needs in the<br />

Netherlands: an optimised mixture of objectivity and politics..106<br />

4.1. General features of governmental and financial relations in the<br />

Netherlands ........................................................................................... 107<br />

4.2. Compartments and clusters in the Municipal Fund: volumes and features<br />

of allocation........................................................................................... 111<br />

4.3. Shortcomings observed in the nineties and the search for a new method<br />

............................................................................................................... 116<br />

4.4. The choice for a new method: the analysis of differences ................... 119<br />

4.5. Examples of an improved connection with local needs for expenditure<br />

using the analysis of differences ........................................................... 125<br />

4.6. The institutional environment of the Municipal Fund.......................... 129<br />

5. Assessing regional and local government expenditure needs in<br />

Italy. Small achievements and large prospective issues. .........131<br />

5.1. Introduction .......................................................................................... 131<br />

5.2. The present system of intergovernmental grants in Italy ..................... 132<br />

5.3. Towards a new system ......................................................................... 143<br />

5.4. Conclusions .......................................................................................... 148<br />

6. Equalisation of the settlement pattern and size of the<br />

municipality in Norway ..................................................156<br />

6.1. Why make allowances to the settlement pattern and the size of the<br />

municipality in the Grant scheme........................................................ 156<br />

6.2. Settlement pattern – criteria ................................................................. 158<br />

6.3. From criteria to weights in the calculation matrix................................ 162<br />

6.4. Size of the municipality – compensation for diseconomies of scale.... 163<br />

6.5. Current work ........................................................................................ 165<br />

7. <strong>Measuring</strong> expenditure needs: Japan’s experiences...........166<br />

7.1. Decentralisation in Japan...................................................................... 166<br />

7.2. Intergovernmental relations.................................................................. 169<br />

7.3. Fiscal equalisation ................................................................................ 171<br />

7.4. <strong>Expenditure</strong> needs ................................................................................ 175<br />

4


7.5. Identifying the issues............................................................................ 178<br />

7.6. Revitalising the equalisation system .................................................... 183<br />

8. Annual fluctuations in the cost equalisation and budget<br />

stability for the municipalities in Sweden ...........................188<br />

8.1. Why is local government equalisation necessary ............................... 188<br />

8.2. New system for local government financial equalisation 2005............ 193<br />

8.3. Calculation of grant or charge in cost equalisation .............................. 201<br />

8.4. Annual update of cost equalisation ...................................................... 209<br />

9. <strong>Expenditure</strong> needs equalisation at the local level: methods and<br />

practice .....................................................................213<br />

9.1. Introduction .......................................................................................... 213<br />

9.2. The rationale for expenditure needs equalisation................................. 214<br />

9.3. Methods of needs assessment............................................................... 221<br />

9.4. <strong>Needs</strong> variables and disability factors .................................................. 227<br />

9.5. Funding the expenditure needs equalisation......................................... 233<br />

9.6. Conclusion............................................................................................ 236<br />

10. Issues in fiscal needs equalisation in Germany ................241<br />

10.1. Introduction ........................................................................................ 241<br />

10.2. Intergovernmental fiscal relations in Germany .................................. 243<br />

10.3. The system of fiscal equalization at the local level in Germany ........ 246<br />

10.5. Conclusion.......................................................................................... 265<br />

11. Fiscal equalisation in Australia: Some technical issues .....269<br />

11.1. Introduction ........................................................................................ 269<br />

11.2. The organisation of equalisation ........................................................ 270<br />

11.3. The role of politics ............................................................................. 273<br />

11.4. Are expenditure needs an objective reality....................................... 275<br />

11.6. Main drivers of equalisation............................................................... 280<br />

11.7. Capital spending................................................................................. 286<br />

11.8. Over what time frame should equalisation be achieved ................... 288<br />

11.9. Data issues.......................................................................................... 289<br />

11.10. The use of assessments..................................................................... 290<br />

List of Contributors (CVs)................................................293<br />

5


i.<br />

i.Foreword<br />

Intergovernmental fiscal relations reflect various aspects of a<br />

country's governmental system, which is not only influenced by<br />

economic conditions but also by politics and institutions. These<br />

also make up an evolving system, continuously responding to the<br />

changes in economic, demographic and political variables. Due to<br />

their complicated nature, few countries seem to be satisfied with<br />

their own system of intergovernmental fiscal relations. Most<br />

countries therefore have the desire to learn from other countries'<br />

experiences. However, understanding the intergovernmental fiscal<br />

relations of other countries is not an easy task since detailed<br />

information is often unavailable to outside observers. Also,<br />

countries with mature decentralisation have a system of<br />

intergovernmental fiscal relations that cannot be easily mimicked<br />

by countries in the early stages of decentralisation. Even with<br />

these difficulties, an international comparison of<br />

intergovernmental fiscal relations provides a good learning<br />

opportunity as long as other countries' experiences are understood<br />

in a guided manner. Therefore, combining the insight of academics<br />

and the experiences of practitioners is a valuable way to get<br />

maximum benefit from an international comparative analysis.<br />

In the workshop held in Copenhagen in 2007 and organised jointly<br />

by the Korea Institute of Public Finance (KIPF) and the Danish<br />

Ministry of Welfare, a rare opportunity was created when many<br />

renowned academics and experienced practitioners were gathered.<br />

The specific topic chosen for the workshop was "measuring<br />

expenditure needs." This issue is an important subject in the field<br />

of intergovernmental fiscal relations, but it is usually dealt with<br />

6


within the government. In the workshop, both academics and<br />

practitioners had the opportunity to exchange diverse views on the<br />

subject, which has led to the publication of this volume.<br />

The KIPF has long been devoted to the research on<br />

intergovernmental fiscal relations in Korea. It is also an active<br />

member of the OECD Fiscal Network on Intergovernmental Fiscal<br />

Relations across Levels of <strong>Government</strong>. In 2007, the KIPF joined<br />

the Danish Ministry of Welfare to establish an Asia-Europe joint<br />

research cooperation. Since Korea has a relatively short history of<br />

decentralisation, and many other countries in Asia are at early<br />

stages of decentralisation, this joint research venture will be<br />

beneficial not only to OECD member countries but also to those<br />

non-OECD member countries which are in the process of<br />

decentralisation.<br />

Yun Hi Won<br />

President<br />

Korea Institute of Public Finance<br />

7


ii.<br />

ii.Opening address<br />

On behalf of the Danish ministry in charge of handling<br />

equalisation and block grant systems of local governments and as<br />

the local Copenhagen host of the workshop I want to make a few<br />

introductory comments on the subject of the two-day seminar:<br />

measuring local government expenditure needs.<br />

It is natural for me to take the Danish situation as my starting<br />

point. Denmark is a small country with a rather homogeneous<br />

population – and consequently there is also a common perception<br />

of many aspects of society, not least when it comes to demand and<br />

supply of local public goods. Still, ideas and ways of doing things<br />

vary between individuals as well as localities. First, this goes for<br />

tastes and preferences. It is an important policy goal for the<br />

government to create welfare options and alternatives for Danish<br />

citizens in various dimensions: across local governments, across<br />

institutions and across public and private service providers<br />

financed by or via the public sector. Second, there is a variance in<br />

the methods of welfare services production. Such methods should<br />

be continuously developed and evaluated over time in order to find<br />

the most efficient and cost-effective ways of supplying public<br />

services.<br />

All in all, those factors are significant when explaining why the<br />

Danish government - governing a unitary country - stresses the<br />

merits of a decentralised public sector. <strong>Local</strong> governments include<br />

regions and municipalities, which are in charge of most public<br />

sector tasks and are financed primarily by block grants and – in<br />

the case of municipalities - local taxes. With a decentralised public<br />

8


sector with an important element of local self-governing and<br />

depending heavily on block grants and own taxes, it is natural that<br />

questions about equalisation and the measurement of expenditure<br />

needs have been important in the Danish context, where the aim is<br />

to ensure equalisation of the possibilities open to local authorities.<br />

Furthermore, the Danish public sector has just undergone a major<br />

reform that included all levels of government and tasks as well as<br />

finances. The reform entailed the most profound changes of the<br />

public sector since the 1970s and included both an amalgamation<br />

reform of local governments, a change in the division of tasks<br />

between levels of government, as well as a reform of the local<br />

government financial system. It was no surprise that this also<br />

brought the equalisation criteria on the scene.<br />

This situation may bear similarities to many other OECD- and<br />

developing countries. The public sector organisation differs from<br />

country to country – or even from state to state within a country -<br />

and the development is certainly not synchronised, but may<br />

happen gradually, stepwise or via ‘reforms’. Nevertheless, I assume<br />

that equalisation systems and expenditure needs are very relevant<br />

and important subjects to discuss and make decisions about in<br />

many countries and governments. So, the next question is: do such<br />

decisions have to be taken locally solely based on national factors -<br />

or can we learn from international comparisons and exchange of<br />

experiences<br />

In my opinion, history demonstrates that participants in this<br />

agenda have been confident in learning from international<br />

experiences. Although there has been a need to cope with<br />

organisational differences, the quite technical and quantitative<br />

character of the problems – at least on the surface - may have<br />

helped the ‘translation’ from one system to another. The<br />

organisations represented at this workshop – the IMF, the OECD<br />

and the World Bank – have all made contributions across countries<br />

with respect to equalisation and measuring of expenditure needs,<br />

and the Council of Europe has also done work in this field. I am<br />

also proud to remind you that a seminar was actually held here in<br />

9


Copenhagen nearly 30 years ago (December 1979 1 ), which had the<br />

same agenda: measuring expenditure needs. Like today, both<br />

practitioners from the government level and theoreticians from<br />

universities and similar institutions attended the seminar. As a<br />

matter of fact, even a couple of attendees from the 1979 workshop<br />

are here today – Mr Jørgen Lotz and Mr Giorgio Brosio; this seems<br />

to demonstrate that equalisation and needs analysis can be a longstanding<br />

relationship!<br />

However, it must be recognised that the agenda has developed<br />

since 1979. Production technologies for welfare services have<br />

developed and become more diversified, new areas of the (local)<br />

public sector have appeared on the scene or have become more<br />

important and/or sophisticated, such as environment and health,<br />

whereas others may have diminished in relative importance. On<br />

the technical measurement dimension, the sources of data have<br />

expanded, for example the geographical information systems (GIS),<br />

and statistical methods have become more advanced. Finally,<br />

computational resources have grown vastly. Consequently, we<br />

have to consider what this entails for the measurement and<br />

presentation of expenditure needs.<br />

On the other hand, it is my impression that many of the challenges<br />

and dilemmas from earlier years still apply – to the discussion of<br />

equalisation systems in general, but also to expenditure needs in<br />

particular. First, this goes for the classic dilemma of fairness that<br />

rests upon an elaborate and detailed, but usually also complex,<br />

measurement of numerous dimensions of needs, versus simplicity<br />

and transparency, where only fewer and simpler criteria represent<br />

the expenditure needs. Second, we have the dilemma of creating a<br />

stable system with only minor changes from one year to the next<br />

versus regular and timely updating of measures to keep the system<br />

as close to actual needs as possible. Third, it is often a serious<br />

challenge to sort out service levels and other supply-side factors<br />

from the citizens’ ‘genuine’ needs. Fourth and finally, the effort of<br />

having a system with appropriate incentives but still getting as<br />

close as possible to real needs may also make the task of<br />

1<br />

”<strong>Measuring</strong> <strong>Local</strong> <strong>Government</strong> <strong>Expenditure</strong> <strong>Needs</strong>: The Copenhagen Workshop”,<br />

published by OECD, Urban management studies, 1981.<br />

10


measuring expenditure needs troublesome. Or – to put it another<br />

way – we still have an exciting and challenging agenda.<br />

Taking into account also the potential distributional effects of<br />

equalisation-systems, it should be no secret that changing<br />

equalisation systems in general and the expenditure needs element<br />

of them in particular is certainly no walkover for politicians and<br />

ministers. The distributional consequences and local interests,<br />

urban areas versus rural areas, rich versus poor etc. may make<br />

political discussions quite tense and make it difficult to reach a<br />

consensus. In this respect there seems to be some kind of<br />

relationship with discussions about changes in tax systems.<br />

On this background it will, I think, often be of interest to ‘the<br />

government in charge’ to have as many items and questions about<br />

equalisation as possible worked through by experts, academics and<br />

civil servants from the ministries, the local governments’<br />

organisation and research institutions. This may be especially<br />

relevant in the case of measuring expenditure needs. The<br />

possibility of doing so has been the governing idea behind this<br />

workshop, and on behalf of the Ministry I hope that this workshop<br />

will be useful in this respect - to the individual attendees, to the<br />

organisations and to others that may have the opportunity and<br />

wish to learn about the results. I would also like to stress that in<br />

my understanding we are speaking at this workshop on the<br />

background of our own personal capacities, experiences and<br />

competences – and not on behalf of our institutions.<br />

Niels Jørgen Mau<br />

Deputy Permanent Secretary<br />

Danish Ministry of <strong>Social</strong> Welfare<br />

11


iii.<br />

iii.Introduction<br />

Junghun Kim and Jorgen Lotz 2<br />

0. The book<br />

This book includes a selection of slightly revised papers that were<br />

presented at the Copenhagen Workshop 2007 on <strong>Measuring</strong> <strong>Local</strong><br />

<strong>Government</strong> <strong>Expenditure</strong> <strong>Needs</strong>. All the papers presented for the<br />

conference are available on the homepage of the Danish Ministry of<br />

Welfare and the Korean Institute of Public Finance 3 .<br />

The participants in the workshop represented national<br />

governments, international organisations and the academic world.<br />

This combination of academics and practitioners helped, as it is so<br />

often the case, highlight the political and administrative feasibility<br />

of theoretical models.<br />

The workshop was organised almost 29 years after a workshop on<br />

the subject of measuring expenditure needs held in Copenhagen in<br />

1978. The purpose of the 2007 workshop was to take stock of the<br />

progress made in the methods of measuring local expenditure<br />

needs over the past 29 years, as well as to survey the present-day<br />

practical application of such methods.<br />

However, as those who read the previous volume 4 will realise,<br />

progress made in the theoretical basis since then has not been<br />

2<br />

This introduction is our interpretation of the authors - besides some of our own<br />

thoughts - and we are solely responsible for any possible misrepresentations."<br />

3 http://im.dk/im/site.aspxp=3417 and http://kipf.re.kr/Copenhagen2007<br />

4<br />

OECD (1981)<br />

12


impressive. Bibliographies are much longer today 5 , and some<br />

things are clearer than at that time, but nothing like giant leaps in<br />

new insights have been made, and much of it related to emerging<br />

economies. What has changed, however, is the fact that many more<br />

countries now equalise differences in local expenditure needs.<br />

The 2007 workshop left several unanswered questions. We still<br />

have no clear answer how to evaluate a successful system. The<br />

very different attempts by academics and international<br />

organisations to classify countries’ actual practice suggest that<br />

there is no agreement on this issue.<br />

The workshop was also seen as a supplement to the recent OECD<br />

survey on equalisation 6 .<br />

In this introduction, we aim at identifying general and crosscutting<br />

issues from the papers in the book. But it should be noted<br />

that in addition to the discussion of these issues, several papers<br />

include valuable descriptions of the various national systems of<br />

equalisation. Some even conclude by offering well-argued proposals<br />

for changes in national legislations.<br />

1. Is equalisation of needs necessary<br />

1.1. Are the differences in needs worth the bother of equalisation<br />

One issue we had not really imagined would come up at the<br />

conference, but which actually did, is whether the estimation of<br />

expenditure needs is worth the bother Does it involve more<br />

trouble than equalisation of needs is worth Shah raises a<br />

provocative and fundamental question in this regard and argues<br />

that it is better to avoid fiscal needs equalisation in the general<br />

grants system for reasons of simplicity, transparency and<br />

accountability. Instead he proposes the use of specific purpose<br />

activity-based grants for needs equalisation. This proposal merits<br />

more discussion, but almost all countries participating in the<br />

workshop adopt needs equalisation systems in practice.<br />

5 Throughout this introduction we refer to a few of the more recent surveys on the<br />

issue.<br />

6 OECD (2007)<br />

13


Lennart Tingvall presents in simple terms the classical argument<br />

for equalisation of needs. He shows how, in Sweden, the difference<br />

between the highest and lowest local authority spending on care<br />

for the elderly accounts for a 9 percentage point difference in local<br />

income taxes. And Nobuko Mochida describes how the share of<br />

population over the age of 65 varies across prefectures from 25 to<br />

13 per cent.<br />

Based on Danish data, Niels Jorgen Mau examines two other<br />

possible arguments against needs equalisation and concludes that<br />

such arguments are not supported in his empirical analysis:<br />

Firstly, he examines whether the socio-economic indicators of<br />

needs (like low-income families, unemployed etc.) are already<br />

accounted for in the variation in tax base per inhabitant, and he<br />

finds a significant but modestly negative relationship between<br />

socio-economic stress factors and tax base.<br />

Secondly, Mau examines whether needs caused by a relatively<br />

large number of elderly citizens may be systematically<br />

compensated for by a relatively smaller number of school-age<br />

children. He finds a significant, negative relationship between the<br />

relative numbers of elderly and children (on average, if the share of<br />

elderly people is 10 percentage points higher, the share of children<br />

is 10 per cent lower). However, the weights are not similar. The<br />

elderly are far more expensive than schoolchildren. Furthermore,<br />

since these are average shares, there are also municipalities that<br />

do not fit into this picture and qualify for equalisation grants on<br />

both accounts.<br />

However, in the case of Australia, John Spasojevic describes how<br />

fiendishly difficult it is to ensure that there is no double-counting<br />

in the calculations of weights for the needs criteria. Despite this<br />

kind of shortcoming, Australia still maintains the needs<br />

equalisation element in its system of intergovernmental transfers.<br />

1.2. The relevance of equalisation of needs differs between<br />

countries<br />

While most countries do adopt needs equalisation, the conditions<br />

that make needs equalisation desirable do not exist equally in all<br />

countries.<br />

14


Whether needs equalisation is desirable may in particular depend<br />

on the following, often interrelated, criteria:<br />

1. functions delegated to local authorities.<br />

2. average size of local authorities.<br />

3. geographical size and diversity of countries.<br />

4. urban/rural divide.<br />

First, traditional local functions like environmental services and<br />

utilities may not call for equalisation. They are basically well<br />

suited for user-financing through fees and charges, or via local<br />

taxes. But this is not the case when local functions effect<br />

redistribution-like spending on schools or elderly care. <strong>Local</strong><br />

functions like those are bound to be met with strong national<br />

preferences for minimum standards, which means that<br />

equalisation is called for.<br />

Second, large local units may have a more uniform composition of<br />

population and of other characteristics that create spending needs<br />

than is the case of countries with smaller local units. In his paper,<br />

Mau gives an empirical illustration of this effect.<br />

A federal country like Canada, with large provinces, only<br />

equalises differences in tax capacity. Tingvall argued at the<br />

workshop that needs equalisation is not needed in Canada because<br />

differences in needs are relatively unimportant. But there were<br />

other explanations. Anwar Shah, supported by Bernard<br />

Dafflon/Peter Mischler, argues that the explanation is that Canada<br />

finds that needs equalisation would reduce the transparency and<br />

objectivity of the system.<br />

Third, related to this issue is the population size and geographical<br />

variation of a country. Countries with small populations may span<br />

over smaller variations in living conditions than larger countries.<br />

Variation in Korea is probably larger, and natural conditions more<br />

varied, than in Denmark and the Netherlands (though the same<br />

may not be the case when comparing Korea with Norway).<br />

Fourth, very different regional economic conditions in rural and<br />

urban regions often create difficulties for equalisation. When such<br />

15


differences are very large, equalisation is often implemented by the<br />

introduction of specific, regional types of grant.<br />

1.3. Does equalisation of needs result in undesirable incentives<br />

Some hold the view that equalisation of expenditure needs results<br />

in undesirable incentives. The OECD (2007) has expressed some<br />

concern that cost equalisation can give sub-central governments<br />

the leeway to influence expenditure needs and thus lead to inflated<br />

equalisation payments.<br />

At the workshop, only few expressions of general concern over<br />

possible undesirable incentives were put forward. But<br />

Dafflon/Mischler seem to support the OECD view. They argue that<br />

with full equalisation, there is no incentive for local authorities to<br />

adapt good long-run policies, because they would know that the<br />

resulting differences will be automatically equalised.<br />

Mochida sees a risk that formulation of local expenditure needs is<br />

influenced by other factors – like the need to stimulate the<br />

economy - and in this way will result in local expenditure<br />

pressures.<br />

Kim argues that needs are difficult to measure, and inaccurately<br />

measured needs direct funds to authorities that do not need them,<br />

which may result in an inefficient behaviour of local governments.<br />

Spasojevic explains that some states have maderecent proposals to<br />

reduce the degree of equalisation to “improve incentives for states<br />

to pursue economic efficiency and build their own tax bases”, but<br />

that this has not convinced the government to change policy.<br />

2. How do we measure expenditure needs<br />

2.1. <strong>Measuring</strong> expenditure needs: an empirical or a political<br />

process<br />

The call for the Copenhagen conference described the politically<br />

important distinction between local expenditure needs seen as a<br />

central, politically (normative) determined standard, or as an<br />

empirically defined concept based on local budget data, or a<br />

mixture of the two.<br />

16


In the 1978 workshop, a mixture was the rule: empirical regression<br />

equations were not used directly in the equalisation schemes, in<br />

the known cases (at that time only England and Denmark applied<br />

regression techniques) results were modified by political<br />

transformations. The same seems to be the case today.<br />

Mau characterises the difference between the two approaches as<br />

either estimating a welfare function empirically or centrally<br />

formulating local government production functions. On the other<br />

hand, Dafflon/Mischler hold that all measures of needs require<br />

knowledge of a production function.<br />

The theoretical answer depends on several assumptions: Are local<br />

council budget decisions really the best expression of the<br />

population’s preferences Or is the centre better at reading<br />

preferences and defining the service standard and what is<br />

permissible And then there are the questions of technical<br />

perfection. Are empirical methods adequate to produce satisfactory<br />

measures of needs And then, from a political point of view, there<br />

is the question whether empirical methods lack transparency.<br />

Vaillancourt and Bird (2006) based their classification on two<br />

criteria: (i) the centre decides what services are to be included and<br />

what “standards” are to be met; or (ii) historical expenditure<br />

patterns are used to design representative expenditure systems,<br />

perhaps by applying regression techniques.<br />

Dafflon/Mischler propose a classification that distinguishes<br />

between whether actual local expenditures are used in the<br />

assessment or not. But they differ from the preceding authors by<br />

including “normatively defined “necessary” expenditures” in the<br />

group that uses actual expenditures.<br />

In the end, the best system must be the one that results in political<br />

acceptance and live the longest. This view is shared by<br />

Vaillancourt/Bird (2006), who conclude that “what is important is<br />

not that the formula used is “correct” but that the results of<br />

applying it are politically viable”. Which method is the best in this<br />

respect is difficult to say, but an indication may be found in the<br />

observation that in practice we found that no country applies<br />

exclusively one or the other method.<br />

17


2.2. The centre formulates the needs (the normative method)<br />

Nobuko Mochida of Japan describes the system in Japan as an<br />

example of a basically central, normative design of the measuring<br />

of needs.<br />

And according to Dutch Hessel Boerboom and Peter Huigsloot,<br />

there seems also to be a clear political element in the application of<br />

empirical estimations in the Netherlands.<br />

Looking closer at the paper by Niels Otter from Germany, it<br />

appears, although this is not explicitly stated, that the German<br />

system of modifications to the number of inhabitants includes<br />

strong normative features (Otter concludes that the measure<br />

appears as a product of “politics and traditions”). Kim also<br />

describes the normative nature of needs measurement in Korea,<br />

which is embedded in the use of modification factors that are used<br />

to calculate expenditure needs. Mau describes how in the end the<br />

weighting of indicators of social needs is determined politically.<br />

But this takes into account more technical considerations about<br />

the single indicators and is combined with the use of regressions to<br />

identify viable indicators to describe the variation in social<br />

conditions.<br />

It may be worth noting that the “normative” method was a<br />

favourite of the Communist states, and they even used the word<br />

“normatives” to describe the weights and criteria for distribution of<br />

grants.<br />

The view that the centre is best at reading the preferences of the<br />

nation is criticised by Shah as being subjective and therefore<br />

potentially controversial.<br />

2.3. The empirical approach: The local councils’ decisions reflect<br />

needs<br />

The empirical approach dominates in other papers. Even in the<br />

Scandinavian countries, where most national welfare services have<br />

been delegated to local supply, quantification of needs are based<br />

empirically on local budgets (as simple average spending or by<br />

means of regression equations), probably because very detailed<br />

statistics are available in those countries.<br />

18


The empirical technique for estimation of needs rests on the<br />

assumption that local authority budgets reveal local preferences.<br />

The variation in local expenditure data is statistically compared to<br />

the variation in selected objective indicators that can be argued to<br />

cause differences in needs. This way it is possible to isolate “needs”<br />

as the variation in the budgets caused by external factors (what<br />

Norwegian Grete Lilleschulstad calls involuntary production<br />

costs). The residual variation is seen as having been caused by<br />

local political choices (voluntary production costs).<br />

But for this approach to result in an objective “true” measure, as<br />

demanded by Shah, it must be assumed that local budgets do<br />

express local needs. <strong>Local</strong> budget needs must not be distorted by<br />

supply side behaviour, such as imperfect political decision making,<br />

or by the effects of previous years’ grant distributions, or by<br />

differences in tax bases.<br />

The estimations also need to take into account possible economies<br />

of scale, and of large city needs as well as rural district needs.<br />

Dafflon/Mischler add the difficult problems of multi-collinearity<br />

and omitted variables bias.<br />

2.4. What difference does the choice of method make<br />

The papers give an interesting idea of how countries have weighted<br />

these two methods, and how the use of empirics and politics is<br />

interwoven in ways that are difficult to understand for outsiders.<br />

The different results of applying one method or the other may not<br />

be expected to be that significant, as it is difficult to believe that<br />

any Parliament would agree on norms without looking at actual<br />

spending.<br />

Rather, the difference may lie in the role of politicians. The<br />

political, normative method seems to be transparent and politically<br />

acceptable while politicians may tend to look at empirical<br />

estimations of needs as a black box that it is difficult to understand<br />

and impossible to influence.<br />

But why is it, in the light of the criticism raised against the<br />

empirical method, that normative standards are not relied on<br />

much more<br />

19


One explanation could be that if the centre defines a “minimum<br />

standard”, it may also feel obliged to make the financing of this<br />

minimum available. It may be more convenient for the centre to<br />

use the statistically calculated needs that represent only some<br />

average standards that have in reality already been financed from<br />

existing local budgets.<br />

Another explanation could be that civil servants have opposed the<br />

method for fear that the more transparent normative method<br />

would shift influence away from them and to the politicians. It is in<br />

the interest of the administrations to have complicated calculations<br />

that are only understood by a few experts. (Valaincourt/Bird (2004)<br />

were not sympathetic to this approach and believed that<br />

“politicians who do not understand who gets what are unlikely to<br />

be long in office”. We wish we could be that optimistic.)<br />

A third explanation could be that the centre realises, as mentioned<br />

by Mau, that it suffers from asymmetric information so that it<br />

lacks the data needed to design a realistic production function.<br />

But Mau describes how there is in Denmark a growing awareness<br />

as to whether statistical methods distinguish properly between<br />

needs and preferences, and he does not preclude that the future<br />

will see changes in the direction of increased use of political,<br />

normative standards in some areas.<br />

2.5. Population as indicator, economies of scale or a U-curve<br />

It is obvious that the size of the population of a sub-central<br />

government must have some relation to its expenditure needs.<br />

A straight relation between spending and population size, so that<br />

average spending is the same for all local authorities, seems not to<br />

require equalisation of needs. But a case for equalisation of needs<br />

arises where, like in Germany, the needs hypothesis is based on<br />

“Brecht’s law” according to which there are higher per capital<br />

needs in larger sub-central authorities. It also arises in cases like<br />

Korea and Norway, where economies of scale (or diseconomies) are<br />

assumed.<br />

20


Dafflon/Mischler use a model framework based on the existence of<br />

a U-curve to analyse the difficulties in differentiating between<br />

economies of scale and the effects of other needs criteria, a concern<br />

shared by Otter.<br />

Otter describes how each of the different German states (länder)<br />

has elaborated on “Brecht’s law” and has introduced modifications<br />

to the measuring of population size that may increase the number<br />

of “notional inhabitants” depending on additional factors (like<br />

members of foreign military, number of pupils etc.). But he also<br />

goes one step further and finds that Brecht’s law should perhaps<br />

really be a U-curve.<br />

Strong economies of scale are assumed in Korea, though the real<br />

story, Kim notes, is perhaps less a belief in economies of scale than<br />

the fact that equalisation is used as an instrument of regional<br />

policy to favour rural authorities. Spasojevic writes that<br />

equalisation also in Australia presupposes the existence of<br />

economies of scale.<br />

Giorgio Brosio describes how the Italian regional system, not based<br />

on objective factors but on historical bargaining, favours the small<br />

and the large authorities, which suggests that Italian authorities<br />

also support the idea of U-shaped local expenditure.<br />

Lilleschulstad finds that “the existence of diseconomies of scale in<br />

municipal service production is well documented”, and Norwegian<br />

equalisation therefore favours the small, thinly populated<br />

authorities.<br />

According to Boerboom/Huigsloot, in the Netherlands a fixed<br />

amount for each authority is included in the needs indicators in<br />

order to compensate for scale effects relating to municipal<br />

administrative organisation. This was also the case in the earlier<br />

Danish measuring of needs, but it was given up because of the<br />

negative incentives for municipal amalgamations. And according to<br />

Mochida, it has also in Japan been argued that equalisation of the<br />

costs related to population size removes the incentives for small<br />

local authorities to merge.<br />

21


Dafflon/Mischler make the relevant observation that for economies<br />

of scale to be taken into account, it needs to be examined whether<br />

diseconomies of scale are caused by unwillingness to cooperate<br />

with neighbouring authorities, or by the inability to do so.<br />

Finally, it is noted that economies of scale do not play an important<br />

role in all countries. Kim points out that if areas such as education,<br />

health and old age care are decentralised, which is not the case in<br />

Korea, there is less reason to expect major scale effects.<br />

2.6. Socio-economic indicators<br />

With respect to Japan and Korea, Mochida and Kim describe two<br />

different kinds of indicator that may be recognised in the systems<br />

of many countries. One is the demographic indicators that measure<br />

the “workload” that local services are faced with, and the other is<br />

the “costs” of providing the services like economies or diseconomies<br />

of scale, fuel costs for heating cold areas etc.<br />

But, as stressed by Mau, needs cannot be described in terms of<br />

population data alone if local authorities are responsible for social<br />

services like social assistance, care for the disabled etc.<br />

Population indicators in Denmark, Mau explains, are<br />

supplemented by socio-economic indicators with normative weights<br />

set out in the legislation (but only after having proved to be<br />

significant in regression analyses), while the weightings of the<br />

population “workload” variables typically result from current<br />

calculations of average spending per person.<br />

In his paper, Tingvall describes how Sweden, like Denmark,<br />

supplements indicators for age structure with indicators for social<br />

structure and population density. And in Norway, as described by<br />

Lilleschultad, socio-economic variables are used as independent<br />

variables for regression estimations.<br />

Finally, Brosio discusses the case of health care equalisation in<br />

Italy, where socio-economic indicators seem to have strong effects<br />

on the quality of health care services. Though spending on health<br />

care is subject to full demographic equalisation, people from the<br />

South still migrate to the North in search of quality. He ascribes<br />

this failure to the lack of indicators for socio-economic and other<br />

factors to supplement the population data. In this regard,<br />

22


Spasojevic argues that the outcome cannot be the same for<br />

indigenous and non-indigenous people because of the different<br />

policy choices made by state governments. Australia does not aim<br />

at equal outcome, but rather at equal capacity to provide.<br />

2.7. Supply side problems<br />

A very troublesome problem with the empirical method is how to<br />

handle the identification issues that arise from missing variables<br />

and interaction between demand and supply sides. In what<br />

Boerboom/Huigsloot refer to as the chicken-egg problem, the<br />

empirical method tends to preserve past injustices because the<br />

historical grants distribution and tax capacity have influenced the<br />

observed spending patterns. Dafflon/Mischler observe that the<br />

effects of previous equalisation need to be removed from<br />

expenditure data before they can be used to estimate needs, “a<br />

challenging process”, they say, “that does not often succeed in<br />

practice”.<br />

Another identification issue arises in the case of income elastic<br />

demand. If high-income local authorities demand higher spending,<br />

this may result in empirical results showing that their needs are<br />

higher. If such effects are not acknowledged as needs that should<br />

be compensated for through equalisation, an income variable is<br />

needed in the regression to identify needs that result from income<br />

elastic demand for the public service.<br />

In the same vein, Mau refers to statistical evidence that a high<br />

density of hospitals and easy access in cities result in more<br />

patients. In statistical estimates, this supply effect will be<br />

described as needs.<br />

Also Otter is concerned about the chicken-and-egg issue in<br />

Germany. He seems to conclude that this problem may mean that<br />

it is impossible to assess expenditure needs by means of budget<br />

figures.<br />

Mau states that in Denmark, the problems of identification are<br />

known, but ignored because of lack of methods to correct for them.<br />

He believes, however, that if the supply-side problems were fully<br />

understood, they could lead to a shift in political decisions from the<br />

23


use of empirical methods to emphasis on normative, political<br />

judgements.<br />

Boerboom/Huigsloot believe that supply-side effects are too serious<br />

to apply regression analysis at all. Instead they introduce what<br />

they describe as a “difference analysis”. This technique combines<br />

discretional scrutiny with statistical comparisons of residuals in<br />

ways they argue permit discretionary exclusion of supply-side<br />

effects.<br />

Lilleschulstad describes an interesting Norwegian procedure. The<br />

dependent variable for primary schools is not spending but rather<br />

the number of teaching hours prescribed by the ministry. This way<br />

a normative element seems to enter the calculation. However, she<br />

adds that a recent Norwegian commission concluded that spending<br />

data were to be preferred because the municipalities get increasing<br />

freedom in how they organise their school systems.<br />

Even if it was technically possible to uncouple the supply-side<br />

effects, all this would still present problems for politicians and<br />

their political advisers. To them the problem is not the supply<br />

effects. Their problem is the political costs of repairing it, to take<br />

money away from some authorities, and to explain to them that<br />

they did not deserve them, as it was aptly expressed by<br />

Vaillancourt and Bird (2006).<br />

3. The relationship to regional policy issues<br />

3.1. The pros and cons of equalisation<br />

The relationship between regional policy and equalisation of needs<br />

has been discussed for many years. Already in the early 1950’s,<br />

A.D. Scott (1950, 1952) argued, in the famous debate with<br />

Buchanan, that equalisation tended to reduce the incentive for<br />

people to move away from deprived regions with bad climate, low<br />

quality soil and underdeveloped infrastructure to regions that<br />

offered more productive employment.<br />

Scott’s views are still shared by many. In the present volume,<br />

Otter cites a German source who argued that pollution, congestion,<br />

or an increasing crime rate serves as an automatic stabiliser of<br />

24


egional diversification that should not be “distorted” by<br />

equalisation. Shah argues in his paper that if it is more costly to<br />

deliver public services in rural areas than in urban areas, it is<br />

inefficient for an equalisation programme to neutralise these cost<br />

differences.<br />

But in some cases, Scott’s way of looking at this issue is too<br />

narrow. For example, net fiscal benefit (NFB) differentials across<br />

local governments may be caused by uneven spatial distribution of<br />

industries like mining or hydroelectric power plants. Where this is<br />

the case, equalisation can improve equity and enhance the<br />

efficiency of the inter-regional allocation of labour by eliminating<br />

NFB differentials. And even in cases where inter-regional<br />

migration is not related to NFB differentials, the policies of most, if<br />

not all, countries participating in the workshop is that service<br />

levels should be reasonable -- or the same -- everywhere.<br />

In practice, many countries support lagging regions with high costs<br />

of public service provision, and the ultimate rationale of such<br />

policies is inter-regional equity. Most of the papers agree with<br />

these views.<br />

Dafflon/Mischler describe how Swiss equalisation favours<br />

mountain regions, but they also argue that the relationship to<br />

population density may be complicated and not easy to catch.<br />

Lilleschulstad and Brosio add that in Norway and Italy,<br />

equalisation has even resulted in service levels being the highest in<br />

the most deprived regions.<br />

Spasojevic describes the aim of the system to be enabling the states<br />

to provide services of the same standard – but, in contrast to<br />

Norway and Italy, with the interesting caveat that this means<br />

similar service to all residents in urban areas, and similar though<br />

different services to all residents in remote areas. He emphasises<br />

that it is not the objective of equalisation in Australia to equalise<br />

the economic potential of its states.<br />

And the equalisation system of the Netherlands supports cities<br />

that have been built on soft soil and are threatening to sink,<br />

although, as Boerboom/Huigsloot explain, only partial<br />

25


compensation is offered “because of the own responsibility of<br />

municipalities for building on this type of soil”.<br />

3.2. Equalisation policies vs. regional policies, do they mix<br />

On this background, it is easy to understand the political<br />

temptation to mix equalisation with regional policy. However, the<br />

workshop offered very different opinions on the wisdom of mixing<br />

these two kinds of policy.<br />

Kim argues in favour of separating regional from equalisation<br />

policies and their instruments. He finds that such a separation of<br />

objectives makes the measure of needs simpler and more<br />

transparent.<br />

According to Tingvall, a recent reform in Sweden supports Kim’s.<br />

Regional policy was removed from the equalisation system, and a<br />

separate “structural grant” has been set up for a limited number of<br />

local authorities with small populations or with serious labour<br />

market problems, mostly municipalities in the central and<br />

northern forest regions.<br />

Lilleschulstad argues, as stated, for diseconomies of scale in<br />

Norway. Equalisation takes this into account, but she<br />

distinguishes between this and the specific regional grants given to<br />

the small, thinly populated Norwegian authorities.<br />

According to Brosio, the constitutional reform in Italy in 2001, not<br />

yet implemented, calls for discontinuation of specific grants<br />

towards regional policy purposes because of the risk of inefficient<br />

budget processes, overspending and regional deficits.<br />

But the opposite view was also argued.<br />

Otter holds the view that the equalisation system needs to<br />

correspond to the state’s regional policy objectives. In Germany, he<br />

writes, “the importance attached to other policy goals like the<br />

regional growth perspective has increased significantly in<br />

Germany over the last 5 years.”<br />

And Mochida includes among those cost factors that need to be<br />

equalised, “differences in costs for promotion of the regional<br />

26


economy”. In Japan, there is also equalisation for “agricultural<br />

promotion” related to the number of farming households.<br />

Also Boerboom/Huigsloot from the Netherlands argue that<br />

equalisation should include regional differences described as crossborder<br />

effects, such as the costs of the use of facilities by citizens<br />

from neighbouring local units. In the Netherlands, regional<br />

indicators like “Frisian island municipalities”, “waterways”,<br />

“condition of soil” etc. are used on needs for spending on roads,<br />

water and sewerage.<br />

Further to the question of the wisdom of mixing regional and<br />

equalisation policies is a, perhaps more technical, question<br />

whether it is possible to design one “just” national needs formula<br />

that covers the smallest rural authorities as well as the largest<br />

cities at the same time<br />

Brosio criticises the fact that the special needs of metropolitan<br />

areas have never been addressed in Italy. Boerboom/Huigsloot<br />

explain that the four largest cities in the Netherlands receive<br />

supplementary grants, but this is because they have been<br />

mandated to perform exclusive functions such as urban renewal.<br />

And Mau describes how a special Danish equalisation scheme is<br />

added to the national scheme for municipalities in the<br />

metropolitan region, and that such a special scheme is also in place<br />

for the poorest (rural) municipalities.<br />

3.3. Differences in local market prices<br />

A related issue is the treatment of regional cost differences that<br />

arise due to different local market conditions. One such issue is<br />

whether differences in the wages of local government employees<br />

should be seen as cost differences and be equalised or as products<br />

of local political decisions that are not to be equalised<br />

John Spasojevic from Australia argues that even when wages are<br />

negotiated locally, there is a need for equalisation based on the<br />

differences in wages in the private sector, because the local<br />

authorities have to compete with them for labour. Some high-wage<br />

Australian states have argued for parallel treatment, arguing that<br />

since they are hit by equalisation on the revenue side they should<br />

equally be compensated for the costs of high public salaries.<br />

27


However, Tingvall reminds us, equalisation should not be<br />

expanded to differences in the inhabitants’ private consumption,<br />

such as housing costs. The latter is a persistent political issue,<br />

raised by the rich suburban authorities, in Denmark, but it has so<br />

far been possible to reject the demand for inclusion of private<br />

property prices in the definition of expenditure needs.<br />

4. Stability of grants over the years<br />

4.1. Stability and reform<br />

<strong>Local</strong> government equalisation in a political perspective is a<br />

difficult issue, and reforms are only made when absolutely<br />

necessary. Many have had the experience that it is impossible for a<br />

government to benefit from such reforms, the local authorities who<br />

lose complain, and those who gain either say nothing - or they<br />

complain that they did not gain enough.<br />

But “reforms” are needed, comprehensive reforms are needed if the<br />

system does not satisfy political objectives, but smoothly running<br />

systems need to be adjusted when the surrounding reality changes.<br />

The paper describes how, when reforms become unavoidable,<br />

politicians want to minimise fluctuations. Theoretical scholars may<br />

not fully grasp how enormous this political desire for minimising<br />

fluctuations in the wake of any changes in equalisation is. This is,<br />

as a matter of fact, an additional supply factor that explains why<br />

equalisation keeps reflecting historical differences so much.<br />

Tingvall describes the problems encountered by Sweden. The<br />

Swedish system has been reformed every 4-5 years, and each time<br />

the reforms have included dampers to prevent the annual change<br />

for any local authority from exceeding 0.08 per cent of the tax base.<br />

If, at the time of the reform, the transitional damping-grants from<br />

past reforms have not yet run their full course, the remaining<br />

uncompensated changes have been included in the new dampers.<br />

This has added to the complexity of the system.<br />

Mau, being an experienced civil servant, go as far as describing<br />

dampers as a positive quality (“pragmatism”) of the equalisation<br />

28


system. In Denmark, legislation is usually not allowed to cause<br />

annual shocks of more than 0.2 per cent of the local tax base.<br />

Dafflon/Mischler seem to suggest that for the sake of stability, the<br />

rules for equalisation should be based on the constitution rather<br />

than annual negotiations on the funding.<br />

4.2. Day-to-day stability and lagged data<br />

A declining population means less people to carry the costs of fixed<br />

assets, and population increases may be costly and require new<br />

infrastructure investments. Many see a need for more stability in<br />

grants on a year-to-year basis. This has in some cases resulted in<br />

the use of multi-year, average indicator data as preferred in actual<br />

cases.<br />

Spasojevic shows how in Australia moving 5-year averages with a<br />

one-year lag are used for equalisation, and without revisiting a<br />

year to correct for past over- or underequalisation, but he adds that<br />

considerations are under way to improve actuality.<br />

In the UK (paper not included in this volume), a comprehensive<br />

system of annual dampers and protectors has been introduced to<br />

such a degree that it was argued that the measure of needs “no<br />

longer can be said to express the local demand”.<br />

Population change is, according to Kim, a dominating concern in<br />

Korea, where the equalisation strongly compensates authorities<br />

with declining population.<br />

Tingvall describes how also Sweden has special rules to<br />

compensate for the local costs of population change, and according<br />

to Mau the same is found in Denmark, though in a more limited<br />

version.<br />

Boerboom/Huigsloot report, on the issue of whether the<br />

unavoidable lags between changes and how they are recorded in<br />

the statistics present a problem, that in the Netherlands such<br />

effects have not been serious in general, but problems have been<br />

identified for fast growing municipalities, and the system is<br />

designed to compensate them. Dafflon/Mischler find that delays in<br />

29


the data and the use of moving averages mean that municipalities<br />

cannot indulge in strategic behaviour.<br />

Tingvall sensibly notes that annual fluctuations describing annual<br />

needs fluctuations are no logical reason for suppressing variation<br />

in the compensatory grants. But politicians may have a different<br />

view on this question.<br />

5. The organisation of the measurement of local expenditure<br />

5.1. Transparency is sometimes lacking<br />

There seems to be general agreement that transparency is<br />

important.<br />

Dafflon/Mischler state that the criteria for needs must be discussed<br />

between the partners, should be explicitly publicised, and should<br />

not be fixed by an expert panel in black-box formulas.<br />

As it appears from the paper by Brosio, the political system in Italy<br />

does not see any need for more openness. It is preferred that grants<br />

are distributed in an obscure way “that seems to please most of the<br />

local government units”. The Italian local government association<br />

is dominated by small authorities and is not trusted as a national<br />

local voice in open negotiations. This does not mean that there are<br />

no consultations taking place in Italy; formal institutions are in<br />

place for dialogue with regions and municipalities. But these<br />

institutions negotiate behind closed doors and have not wanted to<br />

reform the system. And Otter demonstrates how also in German<br />

states the measure of needs is “nothing but a product of political<br />

ratios and old-fashioned traditions”.<br />

In Japan, the voice of local authorities is the Ministry of Interior.<br />

Without that role, Mochida writes, the MOI would be a powerless<br />

ministry. He points out that this brings discussions between local<br />

government and state into the government and gives rise to public<br />

discussions of government internal policies. Mochida notes the<br />

difference to the Nordic countries where the local government<br />

associations are the defenders of the views of local government.<br />

Brosio describes a division of ministerial roles in Italy somewhat<br />

similar to that of Japan.<br />

30


5.2. Organisations of open dialogue<br />

One thing is a desire for transparency. Another thing is how the<br />

dialogue can be organised most efficiently There are several views<br />

on this question, but not much evidence.<br />

No system seems to compare in resources and independence to the<br />

Australian system as described by Spacojevic. The statutory<br />

independent Commonwealth Grant Commission provides advice to<br />

the government on how equalisation could be achieved, and there<br />

are intensive exchanges of views between the Commission and the<br />

Ministerial Council comprising the national and state Treasurers.<br />

The Council will consider the Commission’s recommendations, and<br />

the final decision is made by the national Treasurer.<br />

In other countries the organisation is built on less formalised cooperation<br />

between central ministries and local governments.<br />

Small, technical groups with broad representation seem to be a<br />

preferred and perhaps even relatively cost-efficient part of the<br />

structure.<br />

Several such models are described, the Danish Financial<br />

Committee, the Netherlands’ Fiscal Relations Board, and the<br />

Japanese internal central negotiations with the Ministry of<br />

Interior.<br />

In his paper, Mau argues that openness is important when<br />

applying the empirical method. In Denmark, the coordination has<br />

since the 1970s taken place in the Finance Committee. This is a<br />

group headed by the Ministry of Interior (actually by Mr Mau) and<br />

composed of mid-range civil servants from relevant ministries and<br />

from the local government associations. The latter, of course,<br />

cannot openly support any proposals for changes in equalisation,<br />

but their presence gives brings extra legitimacy to the proposals.<br />

One reason for them to participate, Mau suggests, is their<br />

preference for strengthening equalisation to minimise the risk of<br />

the introduction of “centralistic” earmarked grants.<br />

Boerboom/Huigsloot describe the Dutch system as being not all<br />

that different from the Danish system: there is a Financial<br />

Relations Board that monitors the municipal fund, and an advisory<br />

council with representatives from the local government<br />

31


associations that make proposals to the Ministry of Interior<br />

Affairs. The ministry can choose to present these proposals to<br />

Parliament as legislative bills.<br />

6. Are earmarked grants an alternative to needs equalisation<br />

6.1. The criticism of earmarked grants<br />

In the Council of Europe’s 1986 Charter of <strong>Local</strong> Self-government,<br />

the European ministers agree that earmarked grants should be<br />

avoided as far as possible. And recently the OECD (2007) warned<br />

against earmarked grants because they are administratively<br />

burdensome, reduce sub-central choice, lead to distorted local<br />

budget allocations, and raise considerable efficiency concerns.<br />

The 1980s and 1990s saw a wave of reforms in many countries that<br />

replaced earmarked grants with needs equalisation in the form of<br />

general grants.<br />

Kim describes a reform in Korea in 2005 that consolidated many<br />

earmarked grants into one new grant. The intention was to<br />

transform earmarked grants into block grants, but according to<br />

Kim, the end results so far are not much different from the<br />

previous system of earmarked grants.<br />

Boerboom/Huigsloot describe the change in the Netherlands from<br />

narrowly defined specific grants to wide target-specific grants<br />

using criteria developed along the same lines as the criteria for<br />

general grants. According to them, control of local government<br />

activities in the Netherlands is not done by means of financial<br />

instruments but rather by means of legal instruments.<br />

Mau of the Danish Ministry of Welfare argues in his paper for the<br />

involvement of line ministries in the selection of needs indicators<br />

and their weighting because such cooperation is “a way to convince<br />

them to remain loyal to the general model of finance” instead of<br />

working for their own specific grants.<br />

In Australia a particular issue of coordination between earmarked<br />

grants and equalisation has arisen. When, Spasojevic describes, a<br />

state get earmarked grants, its fiscal capacity will increase and the<br />

32


equalisation grant will be reduced accordingly. Some have<br />

criticised that this is done in a way that allows equalisation to get<br />

primacy over the effects of grants donated by different ministries<br />

for their own allocation purposes.<br />

6.2. Will earmarked grants have a renaissance<br />

However, some participants in the workshop seemed to look more<br />

favourably to a return to earmarked grants.<br />

Shah of the IBRD opened the discussion by declaring that<br />

“simplicity, transparency and local autonomy are preserved by<br />

having fiscal need equalization through service-oriented outputbased<br />

fiscal transfers”. He added that “such transfers would<br />

further enhance citizens’ based accountability for results and<br />

thereby offer a potential for enhancing public confidence in<br />

government operations.”<br />

Shah describes expenditure on primary schools as an example and<br />

proposes to make an “output-based national minimum standards<br />

grant”, a specific grant to be paid to service providers (public and<br />

private schools alike) depending on the number of enrolled<br />

students.<br />

The discussion really concerns whether, from the point of view of<br />

administrative efficiency, it is better to rely on local authority<br />

accountability than on that of the central authorities. If the local<br />

government administration is seen as sufficiently efficient, general<br />

grants with the right indicators for distribution may be seen as the<br />

most efficient way of financing local government.<br />

It is important in this discussion to be aware that if the system is<br />

based on general grants and equalisation, then the legislation is<br />

prepared by the coordinating ministry in charge of the general<br />

grants. But legislations on specific grants, in contrast, are<br />

prepared by the specialised line ministries and may tend to follow<br />

their partial objectives rather than national objectives.<br />

Mochida reports that in Japan the Ministry of Finance advocates<br />

replacing general grants distributed according to a measure of<br />

expenditure needs (the socalled LAT) with specific grants. The<br />

Ministry of Finance’s proposal can be seen as an application of the<br />

33


arguments put forward by Shah, but can also be seen as having<br />

more to do with intergovernmental infighting than with simplicity<br />

and transparency.<br />

6.3. A follow-up workshop<br />

It turned out that the issue of grants design unexpectedly became<br />

one of the most discussed issues at the conference. However, time<br />

at the Copenhagen 2007 seminar did not allow for a serious<br />

discussion of this issue.<br />

Plans are being made by the Korea Institute of Public Finance and<br />

the Danish Ministry of Welfare for a follow-up international<br />

workshop in Copenhagen to take up the issues of grants design.<br />

References<br />

Boadway, R. and Flatters, F., 1982, “Efficiency and Equalization Payments in a<br />

Federal System of <strong>Government</strong>: A Synthesis and Extension of Recent<br />

Results,” Journal of Canadian Economics 15, 613--33.<br />

Boex, Jameson and Martinez-Vazquez, Jorge, 2007, “Designing<br />

Intergovernmental Equalization Transfers with Imperfect Data:<br />

Concepts, Practices, and Lessons,” in Martinez-Vazquez, Jorge and<br />

Searle, Bob (eds.), Fiscal Equalization Challenges in the Design of<br />

Intergovernmental Transfers, Springer, New York, 291--343.<br />

Buchanan, James, 1950, “Federalism and Fiscal Equity,” American Economic<br />

Review 40, 583--99.<br />

Buchanan, James, 1952, “Federal Grants and Resource Allocation,” Journal of<br />

Political Economy 60 208--17.<br />

OECD Urban Management Studies 4, 1981, Cameron, Gordon and Lotz, Jorgen<br />

(eds.): “<strong>Measuring</strong> <strong>Local</strong> <strong>Government</strong> <strong>Expenditure</strong> <strong>Needs</strong>: The<br />

Copenhagen Workshop”.<br />

OECD (2007) Network on Fiscal relations Across Levels of <strong>Government</strong>: Fiscal<br />

Equalisation in OECD Countries.<br />

Scott, Anthony, 1950, “A Note on Grants in Federal Countries,” Economica 17,<br />

416--22.<br />

Scott, Anthony, 1952, “Federal Grants and Resource Allocation,” Journal of<br />

Political Economy 60 534--36.<br />

Bird, Richard and Vaillancourt, Francois, 2007, “<strong>Expenditure</strong>-Based Equalisation<br />

Transfers,” in Martinez-Vazquez, Jorge and Searle, Bob (eds.), Fiscal<br />

Equalization Challenges in the Design of Intergovernmental Transfers,<br />

Springer, New York, 259--289.<br />

34


Chapter 1<br />

Fiscal need equalization: Is it worth doing<br />

Lessons from international practices<br />

Anwar Shah 7<br />

1.1. Introduction<br />

Fiscal equalization transfers are conceptually justified on grounds<br />

of fiscal efficiency and regional fiscal equity. Political imperatives<br />

to have a shared sense of political and economic union pave the<br />

way for instituting such transfers in most large, especially federal,<br />

countries. There is also a common consensus that conceptually,<br />

such transfers should equalize to a specified standard of both fiscal<br />

capacities and fiscal needs. In practice, implementing such a<br />

comprehensive system of equalization transfers represents a<br />

difficult challenge, especially in the difficulties posed by the<br />

objective measurement of expenditure needs. This paper reviews<br />

the conceptual challenges as well as lessons from worldwide<br />

experiences in implementing fiscal need compensation in fiscal<br />

equalization transfers with a view to developing guidance for<br />

practitioners.<br />

The paper concludes that while in theory a strong case for a<br />

comprehensive fiscal equalization can be made, in practice, fiscal<br />

need equalization as part of a comprehensive equalization program<br />

introduces a significant complexity which works against the<br />

7<br />

The author is grateful to Jørgen Lotz, Niels Jørgen Mau and Junghun Kim and<br />

other participants in the workshop in Copenhagen in September 2007 for<br />

comments on an earlier draft of this paper. The author is also grateful to the<br />

Korean Institute of Public Finance for financing his participation in these<br />

meetings. The views expressed in this paper are those of the author alone and<br />

should not be attributed to the World Bank.<br />

35


simplicity, transparency and general acceptability of the program.<br />

This does not imply that fiscal need equalization should be<br />

abandoned in the interest of simplicity and transparency. Instead<br />

simplicity, transparency, and local autonomy are preserved by<br />

having fiscal need equalization through service-oriented (specificpurpose<br />

block transfers), output-based fiscal transfers that impose<br />

no spending requirements on any functions or objects of<br />

expenditures. Such transfers contrast with traditional, earmarked<br />

transfers, which impose conditions on spending for specific purpose<br />

or object of expenditure and subsequent verification/certification of<br />

such expenditures. Such output-based block transfers would<br />

further enhance citizens’ based accountability for results and<br />

thereby offer a potential for enhancing public confidence and trust<br />

in government operations.<br />

The rest of the paper is organized as follows. Section 2 presents a<br />

conceptual justification of equalization transfers. Section 3<br />

discusses practical considerations in designing fiscal equalization<br />

transfers. Special attention is paid to expenditure needs<br />

equalization. Section 4 distills lessons from international practices<br />

in fiscal needs equalization. Section 5 discusses output-based<br />

transfers as a simpler alternative to expenditure need<br />

equalization. A final section presents concluding remarks.<br />

1.2. Bridging the fiscal divide through fiscal equalization<br />

transfers – conceptual considerations<br />

Fiscal equalization transfers are advocated to deal with fiscal<br />

inefficiency and fiscal inequity concerns arising from decentralized<br />

decision-making (Boadway, 2007). These transfers are also<br />

justified on political considerations. Large regional fiscal<br />

disparities can be politically divisive and may even contribute to<br />

the threats of secession (Shankar and Shah 2003). This threat is<br />

quite real: since 1975, about 40 new countries have been created by<br />

the break-up of existing political unions. Fiscal equalization<br />

transfers could potentially forestall such threats and create a sense<br />

of political participation, as demonstrated by the impact of such<br />

transfers on the separatist movement in Quebec, Canada.<br />

Decentralized decision-making results in differential net fiscal<br />

benefits (imputed benefits from public spending minus tax burden)<br />

36


for citizens depending on the fiscal capacities of their place of<br />

residence. This leads to both fiscal inequity and fiscal inefficiency<br />

in resource allocation. Fiscal inequity arises as citizens with<br />

identical incomes are treated differently depending on their place<br />

of residence. Fiscal inefficiency in resource allocation results from<br />

people in their relocation decisions comparing gross income<br />

(private income plus net public sector benefits minus cost of<br />

moving) at new locations; economic efficiency considerations<br />

warrant comparing private income minus moving costs, only<br />

without any regard to public sector benefits. A nation that values<br />

horizontal equity (the equal treatment of all citizens nationwide)<br />

and fiscal efficiency needs to correct the fiscal inequity and<br />

inefficiency that naturally arise in a decentralized government.<br />

Grants from the central government to states and/or local<br />

governments can eliminate these differences in net fiscal benefits if<br />

the transfers depend on the tax capacity of each state relative to<br />

others and on the relative need for and cost of providing public<br />

services. The more decentralized the tax system is, the greater the<br />

need for equalizing transfers. Boadway (2007) notes that<br />

differential net fiscal benefits may also give rise to regional/local<br />

policy choices in response to local preferences. Such differentials,<br />

however, should not be considered for equalization purposes.<br />

The elimination of net fiscal benefits requires a comprehensive<br />

fiscal equalization program that equalizes fiscal capacity (the<br />

ability to raise revenues from own bases using national average<br />

tax rates) to a national average standard and provides<br />

compensation for differential expenditure needs and costs due to<br />

inherent cost disabilities rather than differences that reflect<br />

different policies. Some economists argue that if public sector tax<br />

burdens and service benefits are fully capitalized in property<br />

values, the case for fiscal equalization transfers is weaker, as<br />

residents in rich states pay more for private services and less for<br />

public services and vice versa in poorer states. According to this<br />

view, as argued by Oates (1982), fiscal equalization is a matter of<br />

political taste. This view gained currency at the federal level in the<br />

United States and explains why there is no federal fiscal<br />

equalization program there. In contrast, local fiscal equalization<br />

drives most state assistance to local governments in the USA,<br />

especially school finance.<br />

37


Conceptually, full capitalization requires a small open area with<br />

costless mobility. Most federations and even states in large<br />

countries do not fulfill this condition. As a result, criticism of fiscal<br />

equalization using the capitalization argument may have only<br />

weak empirical support (Shah, 1988a).<br />

1.3. Designing fiscal equalization transfers<br />

In principle, a properly designed fiscal equalization transfers<br />

program corrects distortions that may cause fiscally induced<br />

migration by equalizing net fiscal benefits across states. A<br />

reasonable estimate of the costs and benefits of providing public<br />

services in various states is essential to measure net fiscal benefits.<br />

Measures of differential revenue-raising abilities and the needs<br />

and costs of providing public services in different states must be<br />

developed. Equalization of net fiscal benefits could then be<br />

attempted by adopting a standard of equalization and establishing<br />

the means of financing the needed transfers.<br />

<strong>Measuring</strong> Fiscal Capacity<br />

Estimating fiscal capacity—the ability of governmental units to<br />

raise revenues from their own sources—is conceptually and<br />

empirically difficult. The two most common ways of doing so are<br />

with macroeconomic indicators and the representative tax system.<br />

Macro indicators<br />

Various measures of income and output serve as indicators of the<br />

ability of residents of a state to bear tax burdens. Among the better<br />

known measures are the following:<br />

• State gross domestic product (GDP). State GDP represents<br />

the total value of goods and services produced within a<br />

state. It is an imperfect guide to the ability of a state<br />

government to raise taxes, since a significant portion of<br />

income may accrue to non-resident owners of factors of<br />

production. For example, the Northern Territory has the<br />

highest per capita income in Australia, but it is treated as<br />

the poorest jurisdiction in federal-state fiscal relations.<br />

• State factor income. State factor income includes all<br />

income—capital and labor—earned in the state. It makes no<br />

38


distinction between income earned and income retained by<br />

residents.<br />

• State factor income accruing to residents only. This<br />

measure represents a more useful measure, provided states<br />

are able to tax factor income.<br />

• State personal income. The sum of all income received by<br />

residents of a state is a reasonable measure of the state’s<br />

ability to bear tax burdens. It is an imperfect and partial<br />

measure of the ability to impose tax burdens, however, and<br />

therefore not a satisfactory measure of overall fiscal<br />

capacity.<br />

• Personal disposable income. Personal disposable income<br />

equals personal income minus direct and indirect taxes plus<br />

transfers. This concept is subject to the same limitations<br />

affecting personal income.<br />

In general, macro measures do not reflect the ability of subnational<br />

governments to raise revenues from own sources. Boadway argues<br />

against the use of macro indicators in an equalization formula on<br />

the grounds that a macro formula “ignores the fact that fiscal<br />

inefficiency and fiscal inequity are the products of the actual mix of<br />

taxes chosen by provincial governments” (Boadway, 2002a, 12).<br />

This neglect runs the risk of violating the principles of equalization<br />

itself. A second major difficulty in the use of macro indicators is the<br />

availability of accurate and timely data at subnational levels. Such<br />

data become available only with significant lags, and the accuracy<br />

of such data may be questionable. Use of these data may therefore<br />

invite controversy (see Aubut and Vaillancourt 2001 for a<br />

Canadian illustration of this point). Despite these problems, both<br />

Brazil and India use macro indicators in their federal-state<br />

revenue-sharing programs.<br />

Representative Tax (Revenue) System<br />

The representative tax system approach measures the fiscal<br />

capacity of a state by the revenue that could be raised if the<br />

government employed all of the standard sources at the nationwide<br />

average intensity of use. Estimating equalization entitlements<br />

using the representative tax system requires information on the<br />

tax bases and tax revenues for each state. Fiscal capacity of the<br />

have-not states is brought up to the median, mean, or other norm.<br />

39


Using the mean of all states as a standard, the state equalization<br />

entitlement for a revenue source is determined by the formula:<br />

E = (POP) x<br />

{[(PCTB) i na x i<br />

t<br />

i<br />

x<br />

na<br />

] – [(PCTB) i x x i<br />

t<br />

na<br />

]}<br />

where E i is the equalization entitlement of state x from revenue<br />

source i, POP is population, PCTB i is the per capita tax base of<br />

revenue source i, t i is the national average tax rate of revenue<br />

source i, subscript na is the national average, and subscript x is<br />

state x. The equalization entitlement for a state from a particular<br />

revenue source can be negative, positive, or zero. The total of these<br />

values indicates whether a state receives a positive or negative<br />

entitlement from the interstate revenue-sharing pool. Since data<br />

on major tax bases and tax collections required to implement a<br />

representative tax system are usually published regularly by<br />

various levels of government, the representative tax system does<br />

not impose new data requirements and can be readily implemented<br />

in countries that have decentralized taxing responsibility to sub<br />

national levels, as most transition economies do. Of course,<br />

implementing such a system will not be feasible in countries with<br />

limited tax decentralization (very large vertical fiscal gaps) or poor<br />

tax administration.<br />

<strong>Measuring</strong> <strong>Expenditure</strong> <strong>Needs</strong><br />

The case for fiscal equalization rests on eliminating different net<br />

fiscal benefits across states that give rise to fiscally induced<br />

migration. Such differential net fiscal benefits can arise as a result<br />

of decentralization of taxing authority and decentralized public<br />

expenditures. Differences in the demographic composition of the<br />

population across jurisdictions will result in differential needs for<br />

decentralized public services, such as education, health and social<br />

welfare. Differences in age distribution affect the need for schools,<br />

hospitals, and recreational facilities. Differences in the incidence of<br />

poverty and disease may affect the need for education, training,<br />

health, social services, and transfer payments. Jurisdictions with<br />

higher need factors would have greater need for revenues to<br />

provide comparable levels of public services at comparable levels of<br />

taxation. These need differentials are likely to cause substantial<br />

variations across jurisdictions in the level and mix of public goods<br />

provided, resulting in different net fiscal benefits. A strong case for<br />

40


equalization can be established on grounds of efficiency and equity<br />

to compensate for need differentials that give rise to different net<br />

fiscal benefits.<br />

The fiscal federalism literature treats differential costs as<br />

synonymous with differential needs, but some cost differences may<br />

arise from deliberate policy decisions by sub-national governments<br />

rather than differences in need. Boadway (2004) argues that even<br />

for inherent cost disadvantages, such as differences between urban<br />

and rural areas, the equity advantage of more equal provision<br />

must be weighed against the efficiency costs. If it is more costly to<br />

deliver public services in rural areas than urban areas, it is<br />

inefficient for an equalization program to neutralize these cost<br />

differences. Even in unitary states, the level of public services in<br />

remote, rural, or mountainous areas is usually lower than in more<br />

densely populated urban areas. Under a decentralized fiscal<br />

system, a policy choice must be made about minimum standards,<br />

but there is no justification for providing the same level of services<br />

in remote and urban areas, as, for example, the Australian fiscal<br />

need equalization program does for its Northern Territories.<br />

Instead, as Boadway suggests, one could stratify locations in all<br />

regions by their costs and equalize across regions within<br />

comparable strata. Equalization grants should partially offset only<br />

inherent disabilities, disregarding cost differences that reflect<br />

deliberate policy decisions or differences in the efficiency with<br />

which resources are used.<br />

In practice, expenditure need is more difficult to define and derive<br />

than fiscal capacity. The difficulties include defining an<br />

equalization standard; understanding differences in demographics,<br />

service areas, populations, local needs, and policies; and<br />

understanding strategic behavior of recipient states. Despite these<br />

formidable difficulties, numerous attempts have been made to<br />

measure expenditure need. The approaches can be broadly<br />

classified into three main categories: (a) ad hoc determination of<br />

expenditure needs, (b) representative expenditure system using<br />

direct imputation methods, and (c) the theory-based representative<br />

expenditure system.<br />

(a) Ad hoc determination of expenditure needs uses simple<br />

measures of expenditure needs in general-purpose transfers. The<br />

41


factors used and their relative weights are arbitrarily determined.<br />

Germany uses population size and population density adjustments,<br />

China uses the number of public employees, India uses measures<br />

of backwardness.<br />

The Canadian provinces use simple measures of expenditure need<br />

in their general-purpose transfers to municipalities. These include<br />

population size, population density, population growth factors,<br />

road length, number of dwelling units, location factors (such as<br />

northern location), urbanization factors (primary urban population<br />

and urban/rural class) and social assistance payments (see Shah<br />

1994b). The most sophisticated of these approaches is the one<br />

taken by Saskatchewan, where the standard municipal<br />

expenditure of a class of municipalities is assumed to be a<br />

function of the total population of the class. Regression analysis is<br />

used to derive a graduated standard per capita expenditure table<br />

for municipal governments by population class.<br />

An interesting example of the application of this approach is South<br />

Africa’s use of it in its equitable share transfers to the provinces<br />

(South Africa 2006). The equitable share formula applicable for<br />

2006–08 focuses almost entirely on need factors, with only a 1<br />

percent weight given to negative needs (per capita GDP). The<br />

formula uses the following shares:<br />

• A basic share (14 percent weight) is derived from each<br />

province’s share of the national population.<br />

• An education share (51 per cent) is based on the size of the<br />

school-age population (5–17) and the average number of<br />

learners (grades R–12) enrolled in public ordinary schools<br />

over the past three years.<br />

• A health share (26 per cent) is based on the proportion of<br />

the population with and without access to medical aid.<br />

• An institutional component (5 per cent) is divided equally<br />

among the provinces.<br />

• A poverty component (3 per cent) is based on incidence of<br />

poverty.<br />

• An economic output component (1 per cent) is based on data<br />

on GDP by region.<br />

42


(b) The representative expenditure system using direct imputation<br />

methods (RES-DIM) seeks to create a parallel system to the<br />

representative tax system on the expenditure side. This is done by<br />

dividing sub-national expenditures into various functions,<br />

determining total expenditures by each jurisdiction for each<br />

function, identifying relative need/cost factors, assigning relative<br />

weights using direct imputation methods or regression analysis,<br />

and allocating total expenditures of all jurisdictions on each<br />

function across jurisdictions on the basis of their relative costs and<br />

needs for each function (see Barati and Szalai, 2000, also reported<br />

in Shah, 2007, for a compilation of need factors used in European<br />

countries).<br />

The advantage of this approach is that it obviates the need for the<br />

very elaborate calculations and assumptions needed to quantify<br />

the provision of services at some defined level. It does so by using<br />

the sum of actual total expenditures as the point of departure for<br />

measuring expenditure needs, reducing the problem to one of<br />

allocating total need among sub-national governments on the basis<br />

of selected indicators of need, including proxies for need if desired.<br />

The disadvantage of this approach is that it does not necessarily<br />

exclude expenses incurred by any of the provinces that go beyond<br />

the concept of a “reasonable level of public service.” However, the<br />

approach can be adjusted to exclude identifiable excesses from<br />

total expenditures (for example gold standards for some services or<br />

relatively unaffordable benefits provided by some rich states) in<br />

respect of which needs are to be allocated.<br />

A sophisticated variant of this methodology is used by the<br />

Commonwealth Grants Commission of Australia, which defines<br />

expenditure as the cost of supplying average performance levels for<br />

the existing mix of state-local programs. Relative expenditure<br />

needs are then determined empirically using direct imputation<br />

methods for 41 state-local expenditures. The following hypothetical<br />

example illustrates the treatment of welfare expenditures using a<br />

crude approach similar to that used by the Commonwealth Grants<br />

Commission for establishing expenditure needs under a<br />

representative expenditure system.<br />

Assume that there are 10 states in Grantland, that the unit costs<br />

of welfare are equal in all states, and that needs for welfare vary<br />

43


ased on the percentage of the working-age population that is<br />

unemployed, the percentage of the population that is not of<br />

working age, and the percentage of families with a single parent.<br />

The independent grants commission assigns a 40 percent weight to<br />

the percentage of the working-age population that is unemployed,<br />

a 35 percent weight to the percentage of the population that is not<br />

of working age, and a 25 percent weight to the percentage of<br />

families with a single parent. Assume that expenditures by all<br />

states for welfare total $5 billion and that state A accounts for 4.8<br />

per cent of the 10-state total for the first factor, 3.0 per cent of the<br />

total for the second factor, and 2.2 per cent of the total for the third<br />

factor. State A’s estimated need for a standard level of welfare<br />

expenditure would then equal:<br />

$5 billion x (0.048 x 0.40) + (0.03 x 0.35) + (0.022 x 0.25) = $176<br />

million, or 3.2 per cent of all state expenditures.<br />

Shah (1994a) provides an application of the approach using<br />

provincial-local expenditure functions for Canada and uses<br />

quantitative analysis in selection and weighting of factors for<br />

various expenditure functions.<br />

This approach is highly subjective and therefore potentially<br />

controversial. Recent experience in Australia vividly demonstrates<br />

the problems that arise if such an approach is followed in practice,<br />

as discussed in the following section. Some subjectivity and<br />

imprecision can be alleviated by using quantitative analysis in<br />

choosing factors and weights, as Shah (1994a) suggests.<br />

(c) The theory-based representative expenditure system (RES-TB).<br />

The theory-based representative expenditure attempts to<br />

implement a conceptually desirable view of expenditure needs<br />

equalization objectively, i.e. localities to be compensated for<br />

inherent cost disabilities rather than differences that reflect<br />

different policies and preferences. This means that the influences<br />

of these latter factors must be isolated. By doing so, the<br />

representative expenditure system can be significantly improved.<br />

This is done by using a conceptual framework that embodies an<br />

appropriately defined concept of fiscal need and properly specified<br />

expenditure functions that are estimated using objective<br />

quantitative analysis, as proposed by Shah (1996) for Canada.<br />

44


Under this refined approach, the so-called theory-based<br />

representative expenditure system, the equalization entitlement<br />

from expenditure category i equals the per capita potential<br />

expenditure of state A for category i based on own need factors if it<br />

had national average fiscal capacity minus per capita potential<br />

expenditure of state A on expenditure category i if it had national<br />

average need factors and national average fiscal capacity.<br />

This approach is even more difficult to implement than the less<br />

refined approach of the RES-DIM used in many OECD countries,<br />

but it has the advantage of objectivity, and it enables the analyst<br />

to derive measures based on actual observed behavior rather than<br />

ad-hoc value judgments. The relative weights assigned to various<br />

need factors and their impact on allocation of grant funds are<br />

determined by econometric analysis. Furthermore, this approach<br />

yields both the total pool and the allocation of fiscal need<br />

equalization grants among recipient units. This method requires<br />

specifying determinants for each service category, including<br />

relevant fiscal capacity and public service need variables. A<br />

properly specified regression equation yields quantitative<br />

estimates of the influence each factor has in determining spending<br />

levels of a category of public service. This information can be<br />

analyzed to determine what each state would actually have spent if<br />

it had national average fiscal capacity and but actual need factors.<br />

This then can be compared to the standard expenditure for each<br />

service based upon an evaluation of the same equation for<br />

determining what each state would have spent if it had the<br />

national average fiscal capacity and also national average need<br />

factors. The sum of differences of these two expressions for all<br />

expenditure categories would determine whether or not the state<br />

had more (if sum was positive) or less than the average needs (if<br />

sum was negative) (see Shah 1996 for a Canadian application of<br />

this approach).<br />

The formula for equalization entitlement based on expenditure<br />

classification i for state x could be stated as follows:<br />

EE i x = (POP) x [(PCSE) i x – (PCSE) i na ],<br />

45


where EE i x<br />

is the equalization entitlement for expenditure<br />

classification i for state x, POPx is the population of state x, PCSE i x<br />

is the per capita standardized expenditure by state x on<br />

expenditure classification i (or the estimated amount the state<br />

would have spent to meet actual needs if it had national average<br />

fiscal capacity), and PCSE i na<br />

is the national average per capita<br />

standardized expenditure for classification i. This is the estimated<br />

expenditure for all states, based on national average values of<br />

fiscal capacity and need. The equalization entitlement for a<br />

particular expenditure classification could be positive, negative, or<br />

zero. The total of these entitlements in all expenditure categories is<br />

considered for equalization. This approach has not been used in<br />

practice by any country as of this date in equalization grant<br />

application. This is understandable due to the complexity of this<br />

approach.<br />

A comprehensive system of equalization determines the overall<br />

entitlement of a state by considering its separate entitlements<br />

from the representative tax system and the representative<br />

expenditure system. Only states with positive net entitlements are<br />

eligible for transfers of all or some fraction of the total amount,<br />

with the fraction determined by the central government based on<br />

the availability of funds.<br />

1.4. Practical difficulties in equalizing expenditure needs: a<br />

review of worldwide experiences<br />

A number of countries use expenditure needs equalization in grant<br />

allocation. In this section, we review selective experiences to draw<br />

some general lessons.<br />

Australia<br />

The Commonwealth Grants Commission (CGC) of Australia found<br />

the theory-based representative expenditure system approach<br />

difficult to implement. It opted instead for an alternative<br />

representative expenditure system using direct imputation<br />

methods that simply equalizes what all states on average actually<br />

spend. The use of expenditure need factors is extensive. Several<br />

hundred factors specific to 41 areas of expenditures in three broad<br />

categories are used: (a) scale factors; (b) demographic factors –<br />

these include dispersion, urbanization, social composition and age<br />

46


structure; and (c) environmental factors including physical and<br />

economic factors. Overall approach in assessing expenditure needs<br />

used by the CGC is highly data and subjective judgment intensive.<br />

Continuous refinements over time to accommodate opposing points<br />

of view have led to super complexity and non-transparency.<br />

Further, the approach assumes that costs are independent of<br />

management paradigm and resource use is independent of<br />

incentives. As an example, expenditure need for government<br />

secondary education is determined separately for government and<br />

non-government schools. A mixture of actual and notional<br />

enrolments is used with special weights for diplomatic families.<br />

Student populations from disadvantaged groups are given weights<br />

ranging from 1.1 to 1.7. Year 11 and 12 grades receive 20% upward<br />

adjustment in costs. Factors used in expenditure need<br />

determination include administrative scale, administrative input<br />

costs, service delivery scale, urban influences, humanitarian<br />

refugees, cross border students, vandalism, dispersion, isolation,<br />

school input costs, wages, accommodation, electricity, rural<br />

students and isolation factors. Somewhat different factors and<br />

factor weights are used for government and non-government<br />

schools. If a private school has above average costs, an additional<br />

grant is assessed, although the state may or may not finance such<br />

an education.<br />

The Australian system seeks absolute comparability for all 41<br />

state-local services rather than just merit goods (some would<br />

question whether this is worth pursuing). Australia’s<br />

Commonwealth Grants Commission makes these calculations<br />

using broad judgments and sampling services. With the single<br />

exception of the Northern Territory, which has a large aboriginal<br />

population, there is little cross-state variations in the expenditure<br />

needs of the Australian states. A special grant for the Northern<br />

Territory would simplify the Australian program while achieving<br />

its equalization objectives.<br />

Australia’s approach raises several questions. Is equal access to all<br />

services in remote areas desirable at any cost If a rich state<br />

decides to buy limousines for its officials, or make higher welfare<br />

payments to its aboriginal population, why should equalization<br />

payments to poorer states go up Such an approach diverts states’<br />

energies to demonstrate that they “need more to do less” or “money<br />

47


does not buy much” as opposed to “doing more with less”, as higher<br />

spending is rewarded and cost-saving in delivering improved<br />

services is discouraged by the equalization grant formula. Such a<br />

system rewards some bad behaviors, including excessive use of<br />

some services by specific groups, tax expenditures by states to<br />

attract capital and labor, and state assumption of contingent and<br />

non-contingent liabilities.<br />

In addition to conceptual difficulties, the Australian program is<br />

plagued with measurement problems. The determinants of<br />

expenditure needs for various expenditure categories are arrived at<br />

based on broad judgments. Arbitrary procedures are used to derive<br />

factor weights and combine various factors into functional forms.<br />

State disabilities stemming from various factors are multiplied.<br />

For highly correlated factors, disabilities are artificially magnified<br />

through double counting and multiplication. Table 1 illustrates<br />

this point where for government secondary education, category<br />

disability is lower than a simple or weighted average of individual<br />

disability factors for rich states and vice versa for poor states.<br />

Under such a program, use of judgment on factors and weights is<br />

inevitable, but such judgments invite controversy and compromise<br />

the credibility of the whole program. The results are often<br />

disappointing. As the commission acknowledges, “given the<br />

number of conceptual and empirical difficulties… and numerous<br />

judgments. different relativities (and grant outcomes) could be just<br />

as valid as those presented [here]”. (Commonwealth Grants<br />

Commission 2000, p.2).<br />

The Australian experience highlights the practical difficulties<br />

associated with implementing fiscal need compensation as part of a<br />

comprehensive fiscal equalization approach (see Shah, 2004, 2007).<br />

State-local transfers in Australia follow the CGC methodology and<br />

are faced with similar measurement issues, although the degree of<br />

difficulty may be considerably less as local governments in<br />

Australia have extremely limited expenditure responsibilities, i.e.<br />

mainly roads and rubbish.<br />

48


Table 1. An example of expenditure need determination in<br />

Australia: secondary education expenditure need factors<br />

<strong>Government</strong> Secondary Education Factors - 1995<br />

Disability Factors NSW Vic Qld WA SA Tas ACT NT<br />

Dispersion 0,9973 0,9921 1,0093 1,0106 0,9972 0,9952 0,9885 1,071<br />

Grade Cost 1,0014 1,0028 0,9966 0,995 0,9992 0,9998 1,0016 0,9979<br />

Input Costs 1,012 0,995 0,986 1,003 0,991 0,99 1,008 1,034<br />

Relevant Population 0,9749 0,8874 1,0983 1,1639 0,9679 1,1422 0,975 1,2226<br />

Administrative Scale 0,9946 0,9946 0,9946 1,0065 1,0105 1,0304 1,0463 1,1139<br />

Service Delivery Scale 0,9922 0,9906 1,0031 1,0153 1,0166 1,038 0,9714 1,1141<br />

Vandalism & Security 1,0023 1,0023 0,9973 0,9973 0,9973 0,9923 0,9923 0,9923<br />

Cross-border 0,9965 1,0001 1,0001 1,0001 1,0001 1,0001 1,066 1,0001<br />

Category Disability 0,9692 0,8658 1,0815 1,1941 0,9772 1,1917 1,044 1,6605<br />

Source: Commonwealth Grants Commission, Australia, 1995-96 Review<br />

Canada<br />

The federal fiscal equalization program is solely focused on fiscal<br />

capacity equalization to a specified standard. A recent <strong>Government</strong><br />

of Canada Panel studied the desirability and feasibility of<br />

introducing expenditure need compensation in the equalization<br />

formula but concluded against its introduction to preserve<br />

transparency and objectivity of the system (see Canada, 2006).<br />

However, to compensate for expenditure needs, equal per capita<br />

block federal transfers are made available to provinces to finance<br />

health and post-secondary education with conditions on minimum<br />

service standards and access and no condition on spending and no<br />

federal oversight on provincial spending on assisted services.<br />

Federal transfers to the three territories nevertheless take<br />

expenditure needs into account in a crude manner by simply<br />

adjusting base year expenditure per capita by the average growth<br />

in provincial spending.<br />

In allocation of their general purpose transfers, the Canadian<br />

provinces primarily focus on fiscal capacity equalization but do<br />

49


take into account only a handful of objective expenditure needs<br />

variables, primarily demographics, especially the urban/rural<br />

composition of the population. This has proven to be manageable<br />

and less controversial. Service specific need variables are<br />

considered in specific purpose transfers (Shah, 1983, 1994).<br />

United Kingdom<br />

General purpose transfers to local governments in the UK are<br />

distributed through the <strong>Local</strong> <strong>Government</strong> Finance Settlement<br />

comprising (a ) revenue support grant ; (b) redistributed business<br />

rates; and (c) Police Grant. The overall system is termed as the<br />

Formula Grant (FG) The FG is calculated as follows:<br />

FG= Relative <strong>Needs</strong> (RN) minus Relative Resources (RR) plus<br />

central per head allocation (CA) and plus/minus stabilization (floor<br />

damping) adjustment (SA)<br />

RN is determined by classifying local expenditures into 6 major<br />

service groups: children’s services, adults’ personal social services,<br />

police, fire, highway maintenance, environmental, protective and<br />

cultural services. Population, social, economic and physical<br />

characteristics of each local authority are used as indicators of<br />

need. An attempt is made to keep these indicators to a manageable<br />

level. Fixed costs of all services are also taken into consideration.<br />

RR is calculated separately for four separate groups (upper tier<br />

services, lower tier services, fire authorities, police authorities) of<br />

authorities by examining a local authority’s (LA) tax base per<br />

capita against minimum LA tax base per capita. CA is determined<br />

residually as the balance of central grant after compensating for<br />

needs. SA adjustment is made to ensure that all LAs receive a<br />

guaranteed minimum increase in grant over the previous year (see<br />

Ponsford, 2007 for further details).<br />

Overall, the UK uses an objective method to assess fiscal capacities<br />

and somewhat subjective methods to assess expenditure needs.<br />

The United States of America<br />

The USA has no formal federal-state and federal-local fiscal<br />

equalization program for reasons outlined earlier, but state<br />

transfers to local governments take both fiscal capacity and<br />

expenditure need into consideration – the latter mainly in specific<br />

50


purpose block transfers such as school finance (see Shah, 2007 for<br />

details).<br />

Other Countries<br />

Denmark, Japan, Netherlands and Switzerland among others take<br />

great care in equalizing expenditure needs in central-local<br />

transfers. However, none of these countries have found a method of<br />

determination that escapes controversy and criticism. In fact,<br />

experience shows that the more complex the approach, the greater<br />

criticism it invites.<br />

In conclusion, there neither is nor will there ever be a holy grail of<br />

expenditure need equalization. All practical methods are by nature<br />

subjective and controversial. Sophisticated and complex methods<br />

have not shown to be superior to simpler determination of<br />

expenditure need equalization through specific-purpose outputbased<br />

fiscal equalization transfers as discussed in the following<br />

section. The opposition to such rough but simple and transparent<br />

justice partly comes from a lack of understanding of these newer<br />

types of transfers and equating them with traditional earmarked<br />

grants. These issues are discussed at length in the following<br />

section.<br />

1.5. A simpler approach to expenditure need equalization: use<br />

of output-based transfers for merit public services<br />

Economic rationales for output-based grants (used interchangeably<br />

with performance-oriented transfers in this chapter) stem from the<br />

emphasis on contract-based management under the new public<br />

management framework and strengthening demand for good<br />

governance by lowering the transactions costs for citizens in<br />

obtaining public services under the new institutional economics<br />

approach. The new public management framework seeks to<br />

strengthen accountability for results by changing the management<br />

paradigm in the public sector from permanent appointments to<br />

contractual appointment and continuation of employment subject<br />

to fulfillment of service delivery contracts. It seeks to create a<br />

competitive service delivery environment by making financing<br />

available on similar conditions to all providers – government and<br />

non-government.<br />

51


The new institutional economics approach argues that<br />

dysfunctional governance in the public sector results from<br />

opportunistic behavior by public officials, as citizens are either not<br />

empowered to hold public officials accountable for their<br />

noncompliance with their mandates and/or for corrupt acts or face<br />

high transactions costs in doing so. In this framework, citizens are<br />

treated as the principals and public officials the agents. The<br />

principals have bounded rationality—they act rationally based on<br />

the incomplete information they have. Acquiring and processing<br />

information about public sector operations is costly. Agents (public<br />

officials) are better informed than principals. Their self-interest<br />

motivates them to withhold information from the public domain, as<br />

releasing such information helps their principals hold them<br />

accountable. This asymmetry of information allows agents to<br />

indulge in opportunistic behaviour which goes unchecked due to<br />

high transactions costs faced by the principals and a lack of or<br />

inadequacy of countervailing institutions to enforce accountable<br />

governance. Results-based accountability through output-based<br />

grants empowers citizens by increasing their information base and<br />

lowering their transaction costs in demanding action.<br />

Output-based transfers link grant finance with service delivery<br />

performance. These transfers place conditions on the results to be<br />

achieved while providing full flexibility in the design of programs<br />

and associated spending levels to achieve those objectives. Such<br />

transfers help restore recipients’ focus on the results-based chain<br />

(figure 1) and the alternate service delivery framework<br />

(competitive framework for public service delivery) to achieve those<br />

results. In order to achieve grant objectives, a public manager in<br />

the recipient government would examine the results-based chain to<br />

determine whether or not program activities are expected to yield<br />

the desired results. To do so, he or she needs to monitor program<br />

activities and inputs, including intermediate inputs (resources<br />

used to produce outputs), outputs (quantity and quality of public<br />

goods and services produced and access to such goods and<br />

services), outcomes (intermediate- to long-run consequences for<br />

consumers/taxpayers of public service provision or progress in<br />

achieving program objectives), impact (program goals or very longterm<br />

consequences of public service provision), and reach (people<br />

who benefit from or are hurt by a program). Such a managerial<br />

focus reinforces joint ownership and accountability of the principal<br />

52


and the agent in achieving shared goals by highlighting terms of<br />

mutual trust. Thus internal and external reporting shifts from the<br />

traditional focus on inputs to a focus on outputs, reach, and<br />

outcomes—in particular, outputs that lead to results. Flexibility in<br />

project definition and implementation is achieved by shifting<br />

emphasis from strict monitoring of inputs to monitoring<br />

performance results and their measurements. Tracking progress<br />

toward expected results is done through indicators that are<br />

negotiated between the provider and the financing agency. This<br />

joint goal setting and reporting helps ensure client satisfaction on<br />

an ongoing basis while building partnership and ownership into<br />

projects (Shah 2005b).<br />

Figure 1. Results chain with an application to education services<br />

Program Inputs Intermediate Outputs<br />

Outcomes Impact Reach Objectives<br />

Improve<br />

quantity,<br />

quality,<br />

and access<br />

to<br />

education<br />

services<br />

Educational<br />

spending by<br />

age, sex,<br />

urban/rural;<br />

spending by<br />

grade level,<br />

teachers,<br />

staff,<br />

facilities,<br />

tools, books,<br />

regulations<br />

Enrollment<br />

s, studentteacher<br />

ratio, class<br />

size<br />

Achieve<br />

ment<br />

scores,<br />

graduati<br />

on rates,<br />

drop-out<br />

rates<br />

Literacy<br />

rates,<br />

supply of<br />

skilled<br />

profession<br />

als<br />

Informed<br />

citizenry,<br />

civic<br />

engagement,<br />

enhanced<br />

international<br />

competitiven<br />

ess<br />

Winners<br />

and losers<br />

from<br />

governme<br />

nt<br />

programs<br />

Source: Shah (2005b)<br />

Output-based grants must have conditions on outputs as opposed<br />

to outcomes, as outcomes are subject to influence by factors beyond<br />

the control of a public manager. Public managers should be held<br />

accountable only for factors under their control. Outcome-based<br />

conditions diffuse enforcement of accountability for results. Since<br />

the grant conditions are concerned with service delivery<br />

performance in terms of quality of output and access, the manager<br />

is free to choose program and inputs to deliver results. To achieve<br />

those results, he or she faces positive incentives by grant<br />

conditions that encourage alternate service delivery mechanisms<br />

53


y contracting out, outsourcing, or simply encouraging competition<br />

among government and non-government providers. This can be<br />

done by establishing a level playing field through at-par financing,<br />

by offering franchises through competitive bidding, or by providing<br />

rewards for performance through benchmarking or yardstick<br />

competition. Such an incentive environment is expected to yield a<br />

management paradigm that emphasizes results-based<br />

accountability to clients with the following common elements:<br />

• Contracts or work program agreements based on prespecified<br />

outputs and performance targets and budgetary<br />

allocations.<br />

• Incentives for replacement of a lifelong rotating<br />

employment with contractual appointments with task<br />

specialization.<br />

• <strong>Local</strong> autonomy and managerial flexibility but<br />

accountability for results to own residents (citizens).<br />

• Redefinition of public sector role as purchaser but not<br />

necessarily provider of public services.<br />

• Adoption of the subsidiarity principle—that is, public sector<br />

decisions made at the level of government closest to the<br />

people, unless a convincing case can be made not to do so.<br />

• Incentives for cost efficiency.<br />

• Incentives for transparency and competitive service<br />

provision.<br />

Under such an accountable governance framework, grant-financed<br />

budget allocations support contracts and work program<br />

agreements that are based on pre-specified outputs and<br />

performance targets. The grant recipient’s flexibility in input<br />

selection—including hiring and firing of personnel and<br />

implementation of programs—is fully respected, but there is strict<br />

accountability for achieving results. The incentive and<br />

accountability regime created by output-based transfers is<br />

expected to create responsive, responsible, and accountable<br />

governance without undermining local autonomy. In contrast,<br />

traditional conditional (earmarked) grants with input<br />

conditionality undermine local autonomy and budgetary flexibility<br />

while reinforcing a culture of opportunism and rent seeking where<br />

such a culture exists (table 2).<br />

54


Table 2. Features of traditional earmarked and output-based<br />

conditional grants<br />

Feature Traditional Earmarked grant Output-based grant<br />

Grant objectives Spending levels Quality and access to<br />

public services<br />

Grant design and<br />

Complex<br />

Simple and transparent<br />

administration<br />

Eligibility<br />

Recipient government<br />

departments/agencies<br />

Recipient government<br />

provides funds to all<br />

government and nongovernment<br />

providers<br />

Conditions<br />

<strong>Expenditure</strong>s on authorized<br />

functions and objects<br />

Outputs -service delivery<br />

results<br />

Allocation criteria<br />

Program or project proposals<br />

approvals with expenditure<br />

details<br />

Demographic data on<br />

service population<br />

(potential clients)<br />

Compliance verification Higher level inspections and<br />

audits<br />

Client feedback and<br />

redress, comparison of<br />

baseline and post-grant<br />

data on quality and access<br />

Penalties<br />

Audit observations on financial<br />

compliance<br />

Public censure, competitive<br />

pressures, voice and exit<br />

options for clients<br />

Managerial flexibility Little or none. No tolerance for<br />

risk and no accountability for<br />

failure.<br />

Rewards for risks but<br />

penalties only for<br />

persistent failure.<br />

<strong>Local</strong> government<br />

Little<br />

Absolute<br />

autonomy and budgetary<br />

flexibility<br />

Transparency Little Absolute<br />

Focus Internal External, competition,<br />

innovation and<br />

benchmarking<br />

Accountability<br />

Hierarchical to higher-level<br />

government, controls on inputs<br />

and process with little or no<br />

concern for results<br />

Results-based, bottom-up,<br />

client-driven; empowers<br />

citizens to hold their<br />

governments to account.<br />

Source: Shah (2007).<br />

55


Even where such a culture of opportunism may not be a concern,<br />

as in most OECD countries, output-based grants create incentive<br />

regimes that promote a results-based accountability culture.<br />

Consider the case in which the national government aims to<br />

improve access to education by the poor and to enhance the quality<br />

of such education. A common approach is to provide grants to<br />

government schools through conditional (earmarked) grants. These<br />

grants specify the type of expenditures eligible for grant financing<br />

(books, computers, teacher aids, and so forth) as well as financial<br />

reporting and audit requirements. Such input conditionality<br />

undermines budgetary autonomy and flexibility without providing<br />

any assurance about the achievement of results. Moreover, in<br />

practice it is difficult to enforce, as there may be significant<br />

opportunities for fungibility of funds. Experience has shown that<br />

there is no one-to-one link between increases in public spending<br />

and improvements in service delivery performance (see Huther,<br />

Roberts, and Shah 1997).<br />

Output-based design of such grants can help achieve accountability<br />

for results. Under this approach, the national government allocates<br />

funds to local governments based on the size of the school-age<br />

population (see Box 1). <strong>Local</strong> governments in turn pass these funds<br />

on to both government and non-government providers based on<br />

school enrollments. Non-government providers are eligible to<br />

receive grant funds if they admit students based on merit and<br />

provide a tuition subsidy to students whose parents cannot afford<br />

the tuition. All providers are expected to improve or at the<br />

minimum maintain baseline achievement scores on standardized<br />

tests, increase graduation rates, and reduce dropout rates. Failure<br />

to do so will invite public censure and in the extreme case cause<br />

grant funds to be discontinued. In the meantime, reputation risks<br />

associated with poor performance may reduce enrollments, thereby<br />

reducing the grant funds received. Schools have full autonomy in<br />

the use of grant funds and are able to retain unused funds.<br />

This kind of grant financing would create an incentive<br />

environment for both government and non-government schools to<br />

compete and excel to retain students and establish reputations for<br />

quality education, as parental choice determines grant financing to<br />

each school. Such an environment is particularly important for<br />

government schools, where staff have lifelong appointments and<br />

56


financing is ensured regardless of performance. Budgetary<br />

flexibility and retention of savings would encourage innovation to<br />

deliver quality education.<br />

Output-based grants thus preserve autonomy, encourage<br />

competition and innovation, and bring strict accountability for<br />

results to residents. This accountability regime is self- enforcing<br />

through consumer (parental in the current example) choice. Table<br />

3 notes a few successful applications of such transfers.<br />

Box 1. Fiscal need compensation through output based transfers<br />

for school finance - an illustrative example<br />

Allocation basis to state/local governments: school age population – population<br />

aged 5-17,<br />

Distribution basis for service providers: Equal per pupil to both government and<br />

non-government schools.<br />

Conditions: Universal access to primary and secondary education. Nongovernment<br />

school access to poor on merit. Improvement in achievement scores<br />

and graduation rates from baseline. No conditions on the use of funds.<br />

Penalties: Public censure, reduction of grant funds and risk of termination with<br />

persistent non-compliance. Grant funds automatically decrease if parents pull<br />

out their children from non-performing school.<br />

Incentives: Grant funds increases automatically as school attracts more<br />

students. Retention of savings for optional use from better management of<br />

resources.<br />

Impact implications: Encourages competition, innovation and accountability to<br />

citizens for improving quality and access. Automatic monitoring and<br />

enforcement provisions through parental choices.<br />

Source: Shah (2007)<br />

57


Table 3. Better practices in simpler and transparent equalization<br />

granted<br />

Grant objective Grant design Examples of<br />

better<br />

practices<br />

Fiscal capacity equalization:<br />

To enable provincial/<strong>Local</strong> fiscal<br />

capacity to provide comparable levels of<br />

public services at comparable burdens<br />

of taxation per capita<br />

<strong>Expenditure</strong> need equalization:<br />

To enable provincial/local governments<br />

to meet minimum or average standards<br />

of merit public services<br />

General non<br />

matching<br />

fiscal<br />

capacity<br />

equalization<br />

transfers<br />

Conditional<br />

non matching<br />

output-based<br />

block<br />

transfers<br />

with<br />

conditions on<br />

standards of<br />

service and<br />

access<br />

Fiscal<br />

equalization<br />

with explicit<br />

standard that<br />

determines<br />

total pool as<br />

well as<br />

allocation<br />

(Canada,<br />

Denmark,<br />

and<br />

Germany)<br />

Road<br />

maintenance<br />

and primary<br />

education<br />

grants<br />

(Indonesia<br />

before 2000)<br />

Education<br />

transfers<br />

(Brazil,<br />

Canada,<br />

Chile,<br />

Colombia)<br />

Health<br />

transfers<br />

(Brazil,<br />

Canada)<br />

Examples of<br />

practices to<br />

avoid<br />

General revenue<br />

sharing with<br />

multiple factors<br />

(Brazil and<br />

India); fiscal<br />

equalization<br />

with a fixed pool<br />

(Australia,<br />

China)<br />

Comprehensive<br />

expenditure<br />

need<br />

equalization as<br />

in Australia<br />

Conditional<br />

transfers with<br />

conditions on<br />

spending alone<br />

(most countries),<br />

pork barrel<br />

transfers (USA<br />

e.g. $200 million<br />

earmark in 2006<br />

for a “bridge to<br />

nowhere” in<br />

Alaska), ad hoc<br />

grants<br />

Source: Shah (2007)<br />

Conditional<br />

capital grants<br />

with<br />

matching<br />

rate that<br />

varies<br />

inversely<br />

with local<br />

fiscal<br />

capacity<br />

Capital grant<br />

for school<br />

construction<br />

(Indonesia<br />

before 2000),<br />

highway<br />

construction<br />

matching<br />

grants to<br />

states and<br />

local<br />

governments<br />

(United<br />

States)<br />

Capital grants<br />

with no<br />

matching and no<br />

future upkeep<br />

requirements<br />

58


1.6.Conclusions regarding the practice of fiscal equalization<br />

Fiscal capacity equalization is relatively straightforward to<br />

comprehend and feasible (with some difficulty) to implement once<br />

a (political) decision is made on the standard of equalization. Fiscal<br />

need equalization is a complex and potentially controversial<br />

proposition, because by its very nature it requires making<br />

subjective judgments and using imprecise analytical methods. An<br />

analytical approach such as regression analysis using historical<br />

data is inappropriate when underlying structures are subject to<br />

change due to technology and other dynamic considerations. Great<br />

care is needed to specify determinants of each service. An ideal<br />

fiscal need equalization system- theory based representative<br />

expenditure system as outlined earlier – is difficult to implement<br />

and therefore for good reasons has not been implemented<br />

anywhere in the world. Instead partial and ad hoc yet complex<br />

equalization is quite commonplace and appears politically popular<br />

yet controversial. Such methods tend to make the system opaque<br />

and in the long run invite citizens’ distrust of government<br />

operations.<br />

This need not be the case. Fiscal need compensation can be more<br />

simply and objectively achieved on a service by service basis for<br />

major local merit services such as education, health, infrastructure<br />

etc. by the use of output-based national minimum standards grants<br />

as done in Canada for health and post-secondary education. Such<br />

grants can use simple and objective service based indicators such<br />

as school age population for school finance, weighted population<br />

with greater weights for infants/children and senior citizens for<br />

health finance etc. Continuity of finance can be assured by<br />

maintaining minimum standards of access and service quality.<br />

Such transfers will preserve local autonomy while enhancing<br />

simplicity, transparency and citizens’ based accountability for<br />

service delivery performance.<br />

References<br />

Aubut, Julie, and François Vaillancourt. 2001. “Using GDP in Equalization<br />

Calculations: Are There Meaningful Measurement Issues” Working<br />

59


Paper, Institute of Intergovernmental Relations, Queen’s University,<br />

Kingston, Ontario, Canada.<br />

Barati, Izabella, and Akos Szalai, 2000. “Fiscal Decentralization in Hungary.”<br />

Centre for Public Affairs Studies, Budapest University of Economic<br />

Sciences, Budapest, Hungary.<br />

Bird, Richard, and Francois Vaillancourt, 2004. <strong>Expenditure</strong>-Based Equalization<br />

Transfers. International Studies Program, Working Paper No. 04-10,<br />

Georgia State University.<br />

Boadway, Robin 2002a. “Revisiting Equalization Again: Representative Tax<br />

System vs. Macro Approaches.” Working Paper, Institute of<br />

Intergovernmental Relations, Queen’s University, Kingston, Ontario,<br />

Canada<br />

———–. 2002b. “The Vertical Fiscal Gap: Conceptions and Misconceptions.<br />

Paper.” Paper presented at the conference “Canadian Fiscal<br />

Arrangements: What Works, What Might Work Better,” Winnipeg,<br />

Manitoba, May 16–17.<br />

––––––. 2004. “The Theory and Practice of Equalization.” CESifo Economic<br />

Studies 50 (1): 211–54<br />

––––––. 2007. Grants in a Federal Economy: A Conceptual Perspective. Chapter<br />

2, pp. 55-74, in Robin Boadway and Anwar Shah, editors, 2007.<br />

Intergovernmental Fiscal. Transfers: Principles and Practice. Washington,<br />

DC: World Bank<br />

Boerboom, Hessel and Peter Huigsloot, 2007. The equalized allocation of local<br />

expenditure needs in the Netherlands.<br />

Robin Boadway and Anwar Shah, editors, 2007. Intergovernmental Fiscal.<br />

Transfers: Principles and Practice. Washington, DC: World Bank<br />

Boadway, Robin, and Anwar Shah. Forthcoming. Fiscal Federalism: Principles and<br />

Practices. New York: Cambridge University Press.<br />

Canada, <strong>Government</strong> of. 2006. Achieving a National Purpose: Putting<br />

Equalization Back on Track. Expert Panel Report on Equalization and<br />

Territorial Formula Financing, Department of Finance. Ottawa:<br />

<strong>Government</strong> of Canada<br />

Filimon, R; T.Romer; and H. Rosenthal. 1982. “Asymmetric Information and<br />

Agenda Control: The Bases of Monopoly Power and Public Spending”.<br />

Journal of Public Economics, 17:51-70.<br />

Gomez-Lobo, Andres, 2002. “Making Water Affordable”. In Brooke, Penelope, and<br />

Suzanne Smith, editors, Contracting for Public Services. 23-<br />

29,Washington, DC: World Bank.<br />

Gonzalez, Pablo. 2005. “The Financing of Education in Chile”. Fund for the Study<br />

of Public Policies, University of Chile, Santiago, Chile.<br />

Gordon, Nora, and Emiliana Vegas. 2004. “Education Finance Equalization,<br />

Spending, Teacher Quality and Student Outcomes: The Case of Brazil’s<br />

FUNDEF.” Education Sector, Human Development Department, Latin<br />

America and the Caribbean Region, World Bank, Washington, DC.<br />

60


Chapter 2<br />

<strong>Expenditure</strong> needs equalisation - reasoning<br />

and organisation of work: the Danish case 8<br />

Niels Jørgen Mau<br />

Summary<br />

The Danish system of equalisation of expenditure needs is seen as<br />

a relevant and integrated part of the general financial system. On<br />

this background the way of measuring expenditure needs is<br />

analysed. In the Danish way of thinking, the normative production<br />

function approach is ruled out in favour of the expenditure<br />

approach, which aims at measuring needs from observed data, but<br />

contains a problem of supply side factors. Statistical analyses are<br />

used to identify needs criteria; however, they only constitute one<br />

step of a longer procedure that involves identification of potential<br />

needs criteria according to certain requirements of objectivity and<br />

choosing the best criteria, taking e.g. negative parameter values<br />

and incentives into account . The procedure helps secure a stable<br />

system that can be “translated” to politicians. The work with<br />

expenditure needs, connected to the way of reasoning, is centred on<br />

the Finance Committee and involves the Ministry of the Interior/the<br />

Ministry of <strong>Social</strong> Welfare, the Ministry of Finance line ministries<br />

and local governments’ organisations. The cooperative model of<br />

organising the work of establishing proposals for new expenditure<br />

needs criteria via the Finance Committee, including both technical<br />

preparation and assessment of alternatives, has been stable for a<br />

8<br />

I have received many relevant and very useful comments from Jørgen Lotz,<br />

formerly of the Ministry of Finance. In the Ministry of Interior and Health I have<br />

had good assistance from Dorte Lemmich Madsen in the form of calculations for<br />

the figures and comments to the paper, as well as from Niels Boye Morving, who<br />

prepared the tables. I am of course solely responsible for both the use of comments<br />

and the data.<br />

61


long period and is evaluated according to three parameters:<br />

reliability, future-orientation and pragmatism.<br />

2.1. Introduction - the case for expenditure needs equalisation<br />

The purpose of this paper is to present the Danish way of<br />

reasoning about expenditure needs and equalisation as well as how<br />

the work on these issues is organised in Denmark, perhaps with<br />

some connection or interaction between reasoning and<br />

organisation. The first section deals with the needs for this kind of<br />

equalisation, section 2 describes the Danish discussion on how to<br />

measure expenditure needs, and finally section 3 deals with the<br />

organisation of work.<br />

2.1.1. Introduction – varying economic conditions across local<br />

governments<br />

<strong>Expenditure</strong> needs vary across municipalities, although not so<br />

much as revenues – see table 1.1.<br />

In Denmark, many municipalities were merged in 2007, and the<br />

number of municipalities was reduced from 271 to 98. Larger<br />

municipalities seem to mean less variation in economic conditions,<br />

which is shown in table 1.1 when comparing the two columns. The<br />

effect of merger on variation is, however, not dramatic.<br />

Table 1.1. Variation in tax bases and expenditure needs, before<br />

and after merger of municipalities (2006-data for 271 and 98<br />

municipalities respectively with the same tasks)<br />

Amounts in DKK Tax base per inhabitant <strong>Expenditure</strong> needs per inhab.<br />

2006 271 mun.s 98 mun.s 271 mun.s 98 mun.s<br />

Minimum 107,047 114,972 32,359 32,829<br />

Maximum 297,638 267,766 43,063 42,410<br />

Range of distribution 190,592 152,794 10,704 9,581<br />

Average (weighted) 142,081 142,081 36,901 36,901<br />

Standard deviation<br />

(weighted)<br />

28,175 27,319 1,537 1,305<br />

62


Because of a smaller variance, the differences in expenditure needs<br />

necessitate fewer transfers than do differences in tax base. 9 This is<br />

illustrated in table 1.2. The size of transfers due to expenditure<br />

needs equalisation can, however, far from be neglected.<br />

Table 1.2. Equalisation transfers due to variation in expenditure<br />

needs and tax bases, 2008<br />

Equalisationpayments,<br />

gross<br />

(+ = reduced<br />

<strong>Expenditure</strong><br />

needs,<br />

mill.DKK<br />

<strong>Expenditure</strong><br />

needs,<br />

per cent of tax<br />

Tax base,<br />

mill. DKK<br />

Tax base,<br />

per cent of<br />

tax base<br />

payment) 1<br />

base<br />

Nationwide equal.<br />

scheme 2 3,046 0.4 6,770 0.8<br />

Metropolitan equal.<br />

Scheme 525 0.2 1,584 0.5<br />

Equal. scheme for<br />

municip’s in<br />

unfavourable ec.sit. 837 0.2 2,330 0.7<br />

Block grants -298 -0.0<br />

Largest loss 6.2 4.5<br />

Largest gain -3.2 -9.1<br />

Total 4,111 0.5 7,811 1.0<br />

1<br />

I.e. reduced payments equalisation system. 2 Including corrections for overequalisation.<br />

Note: computed by comparing two situations – the existing situations versus a<br />

system with no differences in expenditure needs and tax bases respectively per<br />

inhabitant.<br />

2.1.2. <strong>Expenditure</strong> needs criteria<br />

The measure of expenditure needs is based on socalled objective<br />

needs criteria (see further the next section), which are categorised<br />

into:<br />

• Demographic criteria (weight 70 %)<br />

• Socioeconomic criteria (weight 30 %)<br />

The weights of the socioeconomic and demographic criteria<br />

respectively are computed in accordance with the relevant<br />

expenditure ratios. In Denmark, local authorities are responsible<br />

9<br />

The Danish equalisation system and different schemes will not be described<br />

here. To get an impression of the system, see formulas in Mau Pedersen (2007, ch.<br />

IV).<br />

63


for a number of social expenditures like income transfers,<br />

expenditures related to disabled persons etc. The “needs” arising<br />

from these expenditures are allocated to the socioeconomic criteria.<br />

“<strong>Needs</strong>” arising from other local expenditures – usually public<br />

consumption like schools, kindergartens etc. – are assigned to the<br />

demographic criteria.<br />

The weights of the individual demographic criteria equal the<br />

expenditure ratios of the relevant expenditure areas, which are<br />

calculated annually based on the municipal accounts.<br />

The weights of the socioeconomic criteria, however, are fixed by<br />

law. These criteria are meant as “umbrella-variables” to measure<br />

social pressure, and are presented as a socalled ‘socioeconomic<br />

index’.<br />

The 2007 weights for both demographic and socioeconomic criteria<br />

are shown in table 1.3 below. In the Appendix, the weights are<br />

compared with criteria in Norway and The Netherlands.<br />

Table 1.3. Weights of expenditure needs criteria in Denmark<br />

(Nationwide equalisation scheme (“Landsudligningen”), 2007<br />

Demographic exp.n<strong>Needs</strong><br />

criteria<br />

Weight,<br />

per cent<br />

Socioeconomic exp. needs<br />

criteria<br />

Weight,<br />

per cent<br />

Age groups: 20-59 years unemployed ><br />

5%*<br />

5.4<br />

0-6<br />

9.8 25-49 years without<br />

vocational training*<br />

5.2<br />

7-16 21.1 Rented dwellings* 1.5<br />

17-19 1.1 Psychiatric patients* 1.5<br />

20-24<br />

2.0 Families in certain types of<br />

dwellings*<br />

4.5<br />

25-34<br />

5.4 Children in poorly educated<br />

families*<br />

4.5<br />

35-39<br />

2.9 Single more than 65 years<br />

old*<br />

0.7<br />

40-64 11.7 Low income individuals* 3.0<br />

65-74 4.2 Disabled* 1.5<br />

85 years or more 5.6 Immigrants and descendants* 0.7<br />

Commuting time 4.8 Number of living years lost* 0.7<br />

1.4 Decline in population<br />

numbers*<br />

0.7<br />

Total 100.0<br />

Source: Velfærdsministeriet. * = criteria included in the socioeconomic index.<br />

64


2.1.3. <strong>Expenditure</strong> needs equalisation – is it needed<br />

It has been discussed in Denmark from time to time whether the<br />

expenditure needs equalisation, and especially the socioeconomic<br />

element, is really necessary, see Finansieringsudvalget (2004).<br />

A theoretical argument against the necessity of the socioeconomic<br />

element of expenditure needs equalisation is that revenue<br />

equalisation might cover both aspects of equalisation. This would<br />

be the case if socioeconomic needs vary with the tax base. One<br />

hypothesis could be that social pressure is a phenomenon linked to<br />

urban areas, and since urban areas also have the highest taxable<br />

incomes, the socioeconomic equalisation purpose might be served<br />

through moderating the equalisation of revenues. An opposite<br />

hypothesis could be that social problems are correlated with low<br />

income, in which case the revenue equalisation system should be<br />

strengthened with a higher equalisation level. However, as it<br />

appears from figure 1.1, the correlation is negative – but only very<br />

weak.<br />

Figure 1.1. Correlation between socioeconomic expenditure needs<br />

and tax base (data for 2008 equalisation)<br />

65


It has also been argued that demographic differences are minor<br />

and tend to be levelled out since municipalities with e.g. many<br />

children in school and kindergarten age groups (31 per cent of the<br />

demographic criteria, see table 1.3) will be characterised by rather<br />

few elderly persons (about 15 per cent of the criteria). If so,<br />

equalisation of needs is simply unnecessary. This hypothesis has<br />

more, however still rather weak, empirical support from Danish<br />

data, see figure 1.2 below.<br />

Figure 1.2. Correlation between number of children and number of<br />

elderly persons, 2007<br />

2.2. What is the Danish reasoning behind equalisation of<br />

expenditure needs<br />

This section describes the arguments put forward in the Danish<br />

discussion of the measurement of expenditure needs for<br />

equalisation. The sources are partly official reports and<br />

whitepapers.<br />

2.2.1. The demand function<br />

The Danish local government sector is in charge of the main part of<br />

all public sector tasks, and most public sector employees are<br />

66


employed by local governments. In principle, the local governments<br />

have a considerable degree of autonomy in determining the service<br />

level in accordance with citizens’ preferences, although the degree<br />

of freedom is more pronounced in some areas (i.e. kindergartens)<br />

than in others (i.e. schools). Also, local governments are engaged in<br />

social security payments such as cash benefit schemes and social<br />

pensions, but here the autonomy is much less significant, and local<br />

authorities mainly serve as agents of the Central <strong>Government</strong> and<br />

Danish Parliament in carrying out national redistributive policies.<br />

Also in those areas, however, the municipalities typically have<br />

some degree of freedom to choose between different types of social<br />

transfers and arrangements.<br />

All in all it seems natural in a Danish context to consider the<br />

expenditures held by municipalities as a result of the demand from<br />

citizens. On this background, a starting point for expressing<br />

expenditure needs is the demand, D, and the demand function:<br />

10 11<br />

D = PG G (y, pg, X, N ) (1)<br />

Some of these variables may be recognised for equalisation as<br />

expenditure needs that should be equalised – others not.<br />

The price (or cost) factor for inputs used by the municipality, PG,<br />

will normally – or for the most part – be outside the control of the<br />

local authority and should therefore in principle be included in the<br />

calculation of expenditure needs. However, if this factor does not<br />

vary significantly across local governments, it may for reasons of<br />

simplicity be omitted from the equalisation calculations. 12 In the<br />

Danish case, both the lack of high quality local price statistics and<br />

the simplification purposes have resulted in an equalisation<br />

system without price-/cost factors. However, the argument that<br />

prices of inputs to local production are higher in urbanised areas<br />

10<br />

Alternatively, the activities of local governments may be considered in close<br />

connection with Central <strong>Government</strong>’s orders.<br />

11<br />

See Indenrigs- og Sundhedsministeriet (2000) and Mau Pedersen & Møller<br />

(2001) for an empirical test on this equation.<br />

12<br />

The most important cost factor is wages. In Denmark most wage negotiations<br />

take place at the national level, and this suggests that the variation in wages is<br />

limited, see Finansieringsudvalget (2004, ch. 21).<br />

67


has been used to defend a decision to reduce the equalisation<br />

level. 13<br />

The income factor, y (the average level of income and tax base in<br />

the municipality), may be seen as a legitimate expenditure needs<br />

factor that should be equalised (low income connected with high<br />

social expenditures in certain areas), as well as a factor connected<br />

with preferences that should not be included in equalisation (high<br />

income generating high demand of local public goods). Because of<br />

such difficulties, this variable is omitted from the expenditure<br />

needs relation, but an expenditure needs factor connected with low<br />

income in certain households is included directly into the needs<br />

factor, N (see table 1.3 above).<br />

Since prices also determine local demands, the price seen from the<br />

citizen, pG, may also be considered as a possible variable for<br />

equalisation. Of course this price factor is linked to PG, but<br />

moreover the price seen from the citizen depends on the character<br />

of the local public good. If it is – or is close to being - a ‘pure’ public<br />

good, e.g. parks and libraries, the price is higher in sparsely<br />

populated areas than in more urban areas. Therefore, in a system<br />

of full equalisation, this cost or price effect should somehow be<br />

taken into consideration. The Danish solution to this problem has<br />

been – like in the case of y – to include an expression for the<br />

expenditure needs of sparsely populated areas directly into the<br />

expenditure needs as a criterion, see again table 1.3 (the<br />

commuting time criterion).<br />

Concerning preferences, X, they should obviously not be included<br />

in relation to expenditure needs since it is not exactly a need. The<br />

difficult issue is how to distinguish preferences and needs, see<br />

below on ‘objective factors’ supply side effects.<br />

Finally the needs, N, should be included in equalisation.<br />

13<br />

The level was reduced by 5 percentage points, see Indenrigsministeriet (1995)<br />

and Indenrigs- og Sundhedsministeriet (2004, ch.21). It has been shown that this<br />

had almost the same consequences as taking different in prices into account, i.e.<br />

in practice local wages.<br />

68


2.2.2. Norms or expenditures<br />

In principle there are two ways of measuring N, the expenditure<br />

needs. Either the Central <strong>Government</strong> attempts to describe the<br />

production function (or equivalently the cost function) of the local<br />

governments in relevant areas, e.g. schools, nursery homes etc.,<br />

the production function method. Or, alternatively, needs are<br />

assessed on the basis of the observed expenditures (“revealed<br />

preferences”) of individual local governments, the expenditure<br />

method (see further below).<br />

To apply the production function approach, Parliament needs to<br />

define a service level, e.g. how many elderly people of different<br />

categories (very needy, less needy etc.) should receive care and<br />

support for how many hours per week and at which quality level.<br />

Subsequently, the conditions of production, i.e. topographical and<br />

geographical characteristics, need to be taken into account in every<br />

municipality.<br />

The production function method is normative in the sense that<br />

needs are computed based on Parliament’s perception of what<br />

should be the ‘best’ way, i.e. the norm of producing and delivering<br />

services, including how capital, labour and other inputs are<br />

combined in the most effective and efficient manner.<br />

In a Danish context, this production function method has, however,<br />

only seldom been discussed in relation to the general equalisation<br />

system. The method, see Finansieringsudvalget (1978), requires<br />

very good data and knowledge of production possibilities -<br />

information that the Central <strong>Government</strong> in many cases simply<br />

does not have. In addition, the method seems to conflict with a<br />

decentralised public sector where each authority makes its own<br />

calculation and evaluation of the local production circumstances.<br />

The only example from recent years that had some relation with a<br />

production function method was the attempt to calculate<br />

expenditures needs for hospitals also based on – besides age and<br />

sex composition – the composition of diagnoses (DRGs). This way of<br />

calculating expenditure needs for hospitals was, however, never<br />

implemented. In the end, the Organisation of Counties successfully<br />

opposed to the method, expressing its scepticism about the<br />

reliability of the calculations and arguing that the results of<br />

69


certain counties seemed difficult to explain compared to “real world<br />

experience”, see Finansieringsudvalget (1993, s. 2.C). 14<br />

However, this situation may change in the future. If the other<br />

approach to measuring needs, i.e. the expenditure approach, see<br />

below, fails to properly recognise e.g. “needs” and “effectiveness”,<br />

the normative method - using the ‘best’ way of producing the goods<br />

- seems to be an attractive alternative method. Effectiveness and<br />

ways to encourage cost minimisation have been discussed more<br />

intensively in recent years, and this may create a background for<br />

the normative approach (see Finansieringsudvalget (2004, ch. 20).<br />

Also, the local government reform, which reduced the number of<br />

regions to only five, creates problems for stable statistical<br />

estimation in the future and thus difficulties for the<br />

implementation of the expenditure approach. The alternative in<br />

this field may then also be the normative approach.<br />

2.2.3. The expenditure approach and the supply side problem<br />

If we leave out, at least for the time being, the normative method,<br />

the other possibility is to use observed expenditures as a statistical<br />

basis for needs analyses. The purpose is to identify needs criteria<br />

across municipalities.<br />

The consequence of this approach seems to be that all local needs<br />

are to be equalised. At least this has been the case in Denmark. 15<br />

In principle, we should estimate the structural model for supply<br />

and demand, but in practice we have to set up a reduced form<br />

model with independent explanatory right-hand-side variables and<br />

observed expenditures, E, on the left-hand side. The relation (1) is<br />

now expanded into (2) with Z representing supply-side variables:<br />

E = PG (y, pg, X, N, Z) (2)<br />

Concentrating on needs variables, the truncated function is:<br />

14<br />

The DRG-method of calculating the size of grants to regions is, however, applied<br />

when distributing conditional activity-related grants to regions.<br />

15<br />

A proposal to exclude some areas from equalisation of needs (user-financed<br />

utilities, land purchases, administration) was not implemented, see<br />

Finansieringsudvalget (2004, ch. 19).<br />

70


E = f(N) (3)<br />

However, if there is a connection between needs variables and<br />

supply-side conditions, i.e.:<br />

E = g(N, S(N)) (4)<br />

we risk skewed or biased parameter estimates of the needs<br />

variables. If, for instance, we do not equalise expenditure needs<br />

correctly, it is possible that municipalities with high expenditure<br />

needs - because of budget restrictions – need to have a relatively<br />

low service level because of lack of resources and that this would<br />

erroneously be estimated as a low level of needs. 16 Another<br />

example would be the phenomenon ‘supply creating its own<br />

demand’, e.g. an urbanised area with a high ‘density’ of hospitals.<br />

Here the easy accessibility for potential patients ceteris paribus<br />

will result in many patients who will be included in the estimates<br />

as high needs.<br />

Also, differences in effectiveness between local governments, which<br />

are presumably significant, enter into the statistical estimation as<br />

a kind of supply-side phenomenon. This could cause a distortion of<br />

parameter estimations if e.g. low effectiveness and high<br />

expenditure needs are correlated.<br />

In a Danish context, the supply-side conditions are generally<br />

ignored, presumably because of a lack of methods to correct for<br />

them – except for a shift to the production function approach, see<br />

above. The supply-side conditions may be taken into account when<br />

the results of the statistical estimations are interpreted, see below.<br />

But overall, the estimates of needs are probably biased because of<br />

historical supply side effects. In the real world, this may be<br />

justified by the fact that local politicians often experience<br />

expenditure pressure from e.g. individuals with high incomes<br />

demanding them as a kind of “need”.<br />

2.2.4. Implementing the expenditure approach<br />

16<br />

See Finansieringsudvalget (1982, ch. 8) about lags in the adjustment of<br />

expenditures to e.g. growth in population. The problem is also discussed in Lotz<br />

(1987, ch. 7) and Mau Pedersen (2007, ch. V).<br />

71


2.2.4.1. Objective criteria<br />

To be able to carry out estimations of expenditure needs criteria,<br />

we have to pick out a number of measures that will serve as<br />

possible criteria. In the Danish way of reasoning 17 , such potential<br />

criteria have to fulfil certain requirements. The criteria should<br />

express a ‘line of reasoning/causal connection’, should ‘not be<br />

subject to outside influence’, and should ‘be measurable and<br />

computable also by ordinary local authorities’. See also box 1.<br />

Box 1. Three requirements for ‘objective criteria’<br />

1. ‘line of reasoning/causal connection’: in order to have a robust relationship<br />

between criteria and expenditures to be included in the expenditure needs<br />

system, there has to be a reasonable causal relationship between the<br />

criterion and the expenditures. Such arguments are typically easy to<br />

articulate in the case of demographic variables but more difficult for socalled<br />

umbrella variables that represent socio-economic conditions, e.g. the<br />

connection between social expenditures and the character and quality of<br />

dwellings.<br />

2. ‘Not be subject to outside influence’: local governments should not be able<br />

to influence the value of the criterion. Again some of the socio-economic<br />

variables, e.g. number of unemployed, might be somewhat problematic. In<br />

line with this, the so-called local government poverty trap problem,<br />

produced by the revenue and expenditure equalisation in combination, has<br />

created much attention in Denmark, see PLS-Rambøll (2000).<br />

3. ‘Measurability’: local governments should themselves be able to control the<br />

data for their authority. However, with increasing demands of<br />

sophisticated criteria – and improved statistical methods - this<br />

requirement is not always met, e.g. the criterion of accessibility<br />

(commuting time, see table 1.3).<br />

2.2.4.2. Statistical analyses<br />

Having identified a potential list of “objective criteria”, the next<br />

step is to prepare statistical analyses. This has been a tradition in<br />

all government reports about equalisation; see e.g.<br />

Finansieringsudvalget (1998).<br />

<strong>Local</strong> expenditures are categorised into separate “expenditure<br />

blocks” that contain substitutable or overlapping services, and the<br />

individual expenditure blocks are investigated separately.<br />

17<br />

See for example Finansieringsudvalget (1978). See also Council of Europe<br />

(1992).<br />

72


For each expenditure block, a pool of possible expenditure needs<br />

criteria is identified according to the requirements listed in the box<br />

above. To avoid variance linked to the size of municipalities, the<br />

dependent variable is normally divided (or ‘normalised’) by a<br />

demographic measure. Similarly, the independent variables are<br />

calculated per inhabitant (or another variable expressing size) to<br />

avoid size effects. 18<br />

Finally, regression analyses are carried out. Normally, a simple<br />

OLS-regression has been used supplemented by panel data<br />

methods. 19<br />

2.2.4.3. Application<br />

In Denmark, the regression equations are never directly applied to<br />

measure needs as this would result in fluctuations in grants for<br />

each revision. To avoid fluctuations, the regression analysis is only<br />

one step of a longer procedure.<br />

The final stage is the selection of the ‘best’ criteria from the<br />

statistical analyses to find a suitable weight of the criterion.<br />

The weights of the demographic variables, see above, are drawn<br />

directly from local government budgets, and as in accordance with<br />

the law, the weights are automatically updated every year<br />

following the development in the budget data.<br />

There is no automatic updating of the weights of socio-economic<br />

and structural variables. Both the relevance and the weights are<br />

discussed on a regular basis in the Finance Committee<br />

(“Finansieringsudvalget”), see below. For the technical preparation<br />

of the latest reforms of the equalisation system, the main elements<br />

in this process have been: 20<br />

18<br />

See Kabelmann et al (1998).<br />

19<br />

See e.g. Finansieringsudvalget (1998) for the use of panel data methods, and for<br />

the logistic regression method Sundhedsministeriet (1992).<br />

20<br />

Since such considerations are normally not expressed in reports, this is the<br />

author’s assessment. See, however, references below.<br />

73


1) Picking the “best” explanatory variables from the<br />

regressions analyses, also taking into consideration the<br />

parameter values. 21<br />

However, it must be taken into account that the parameter<br />

variables might not be stable from year to year.<br />

2) Exclusion of variables with negative parameter values or<br />

difficulties in “translating” them.<br />

Such variables may not be deemed to be proper expressions for<br />

“need”, or it may be too difficult to get political acceptance of<br />

them. 22<br />

3) Considerations of incentives.<br />

Giving a certain criterion (i.e. number of unemployed), a high<br />

weight may create problems about incentives for local employment<br />

policies.<br />

4) Other relevant considerations.<br />

These may include discussion of supply side factors, e.g. the fact<br />

that certain needs may not be visible or may get low parameter<br />

values via the regression analysis due to limited resources in<br />

certain municipalities. 23 Also, a “problem of outliers” has caused<br />

attention, i.e. the fact that variables passed through step 1)-3) may<br />

nonetheless be poor explanatory criteria when measured against<br />

the expenditure needs of certain municipalities. This discussion<br />

demands micro-based expert knowledge on the part of Committee<br />

members, but it is seldom expressed explicitly in the reports.<br />

Finally, there is a political decision process in Parliament.<br />

21<br />

E.g. Finansieringsudvalget (2005).<br />

22<br />

E.g. Finansieringsudvalget (1998). ’Number of children per household’ showed a<br />

negative but significant relationship with expenditures in kindergartens/child<br />

care. Also ‘number of divorced men’ has shown a significant, positive relationship<br />

with certain socially related expenditures, but this is not used as a criterion (see<br />

the Norwegian system in Appendix.)<br />

23<br />

See a few comments relevant to this issue in Finansieringsudvalget (2005, p.<br />

114).<br />

74


2.3. The organisation of work<br />

2.3.1. Introducing remarks about evaluation<br />

By organisation of work we mean the institutional or<br />

administrative setup that generates proposals for establishing<br />

(change in) expenditure needs. Afterwards, it is up to the<br />

government to make use of such proposals and – if the government<br />

chooses to do so - for Parliament to make decisions about them.<br />

It does not seem obvious in advance what is the ’best’ way of<br />

organising the setup of a proposal for how to establish expenditure<br />

needs for local governments; this also seems to vary to a<br />

considerable extent between countries. 24 Of course there has to be<br />

a political decision process in the end, but what should the<br />

preceding steps be<br />

The question of who should take part in this organisation is related<br />

to the question of how equalisation is handled in relation to the<br />

social welfare function. But who defines this function There are<br />

many candidates or potentially interested partners, but living in a<br />

political environment it would be natural initially to mention the<br />

minister and the government, since the minister in office is the<br />

closest interpreter of the social welfare function. But changes in<br />

equalisation systems are always politically costly, so the<br />

government seeks to take the viewpoints of a broad political<br />

spectrum into account. Also, it is important for the government<br />

that the ‘average’ local government gives its support to – or at least<br />

its acceptance of – the proposals, even though the individual local<br />

governments naturally will often have conflicting interests when it<br />

comes to practice.<br />

Overall, the government demands proposals about expenditure<br />

needs equalisation that have a fair chance of being passed through<br />

Parliament and gain support from the local governments.<br />

To fulfil these objectives, the following ‘standards of quality’ of a<br />

good proposal about expenditure needs are suggested:<br />

24<br />

See e.g. about the separate organisation for equalisation in Australia – the<br />

“Commonwealth Grants Commission” – in the chapter by Spacojevic. In the Dutch<br />

system, see the chapter by Boerboom and Huigsloot, the important so-called<br />

“difference analysis” is carried out by the private consulting firm Cebeon.<br />

75


1. Reliability: the proposal must be reliable. This implies use of all<br />

available and updated data through relevant statistical<br />

methods. It also involves objectivity, i.e. disregarding narrowminded<br />

considerations of serving the benefit of certain favoured<br />

local governments.<br />

2. Future-orientation: the proposal must be forward-looking/focus<br />

on the long term. Since changes in equalisation presumably<br />

will be controversial and involve political “costs”, the proposed<br />

equalisation system should be stable and robust, i.e. not need to<br />

be changed in the short term. It must also include the very<br />

latest information about local government responsibilities and<br />

population forecasts.<br />

3. Pragmatism: the proposal must be well-considered and<br />

pragmatic, i.e. take into account the aspect that the system can<br />

be “explained” to local politicians and the public in general. It<br />

also includes considering distributional consequences, i.e. that<br />

no local government – as far as possible - is “left behind”.<br />

Finally, proposals need to be formulated so that no authority is<br />

exposed to shocks. In Denmark, the ’level of tolerance’ seems to<br />

be negative changes in net revenues from equalisation of 0.2<br />

per cent annually of the tax base .<br />

These elements of a good proposal might potentially be in conflict.<br />

Most obviously, getting too pragmatic may hurt reliability and<br />

future orientation. However, if the proposal does not take<br />

pragmatic considerations into account, it will likely be considered<br />

abstract and theoretical by the political level.<br />

Another dimension of the process of setting up a proposal on<br />

expenditure needs is the different tasks involved in the process.<br />

Three tasks are identified; see also the way of reasoning discussed<br />

in sections 2.4.1 – 2.4.3:<br />

a. Identifying potential criteria: the specialist, but also sensitive,<br />

task of choosing potential criteria that may be candidates<br />

b. Choosing relevant criteria: the statistical and ‘technical’ art of<br />

identifying criteria that have a good and robust correlation<br />

with needs<br />

c. Fixing weights of relevant criteria: the practical task of<br />

proposing weights of criteria based also on distributional<br />

considerations.<br />

76


Overall, this produces a kind of 3x3 matrix that can be used for<br />

evaluating organisation. Reliability may be of general importance,<br />

but when identifying relevant criteria it is possibly more essential<br />

to be future oriented than to be pragmatic. On the other hand,<br />

pragmatism may be rather decisive when it comes to fixing<br />

weights.<br />

2.3.2. Describing the administrative set-up<br />

2.3.2.1. The Finance Committee<br />

Looking back on Danish experiences with the work with<br />

expenditure needs equalisation, the organisation has been<br />

strikingly stable. Since the 1970’s, a committee of civil servants<br />

from ministries and local government organisations has played a<br />

central role in preparing reforms of the equalisation system and –<br />

especially – the expenditure needs part of it. This Finance<br />

Committee (“Finansieringsudvalget”) has produced a large number<br />

of white papers, memorandums and reports. These reports have in<br />

many cases – but certainly not in all - subsequently been followed<br />

up by legislation. Some of the reports were not followed up by<br />

government proposals but did nonetheless play a role in later<br />

discussions about reforms. A few changes/reforms have not been<br />

discussed in the Finance Committee (e.g. the 1995 reform, see also<br />

footnote 17).<br />

2.3.2.2. The Ministry of the Interior/ Ministry of <strong>Social</strong> Welfare 25<br />

Legislation on equalisation and expenditure needs is the<br />

responsibility of the Minister of the Interior, and the MOI provides<br />

the chairman and the secretariat of the committee.<br />

As a rule, work on a new report is initiated on the basis of a<br />

written mandate approved by the minister and the government,<br />

and after the local governments’ organisations have given their<br />

comments. After having been finished, the final report is officially<br />

submitted to the minister for approval. He is, however, updated on<br />

the work in progress on a regular basis.<br />

25<br />

After the elections in autumn 2007, the Ministry of the Interior and Ministry of<br />

<strong>Social</strong> Affairs have been merged into the Ministry of <strong>Social</strong> Welfare.<br />

77


2.3.2.3. The Ministry of Finance<br />

Though the MOI is the minister responsible for local government<br />

organisation and equalisation, no proposal can be advanced without the<br />

support of the MOF. The main interest of the MOF in the work of the<br />

Finance Committee is to secure that local government spending is<br />

within macroeconomically acceptable limits. This means that<br />

stable and robust equalisation models are of interest to this<br />

ministry.<br />

2.3.2.4. The sectoral line ministries<br />

A number of ministries organising specific departments participate<br />

in the work of the committee, i.e. the Ministry of <strong>Social</strong> Affairs, the<br />

Ministry of Labour, the Ministry of Education, The Ministry of<br />

Transportation and the Ministry of Health. These ministries<br />

possess the central government expert information on schools,<br />

kindergartens, hospitals etc. Besides – and not least importantly –<br />

these ministries have the possibility to propose alternative<br />

financial arrangements to block grants combined with<br />

equalisation, in particular different forms of conditional grants, if<br />

they find that the expenditure needs are not mirrored in an<br />

appropriate way in the general financial system. The inclusion of<br />

these ministries in the Finance Committee is also seen as a way to<br />

convince them to remain loyal to the general model of finance.<br />

2.3.2.5. The local governments’ associations<br />

In Denmark two associations that organise local governments<br />

represent all the municipalities and the regions respectively: KL<br />

(National Association of <strong>Local</strong> Authorities, NALA) and Danske<br />

Regioner (Danish Regions).<br />

From the point of view of the associations it is self-evident that<br />

questions of distributions among their members are difficult and<br />

controversial. On the other hand, a fair and well-functioning<br />

equalisation system is a way to avoid different kinds of conditional<br />

grants that are so disliked by local authorities because they are<br />

seen as political centralisation. Finally, the participation allows<br />

local authorities to get their special concerns raised and analysed<br />

by the government.<br />

78


As to the benefits for the Central <strong>Government</strong>, the associations are<br />

assumed to possess the most updated and detailed information on<br />

the financial situation of individual local governments.<br />

The associations can also have their own models analysed, but<br />

usually they focus on offering possible alternative models in the<br />

reports. This must be seen on the backdrop of controversial<br />

distributional effects.<br />

2.3.2.6. The role of external partners<br />

If the role of the Finance Committee has been pronounced, it is<br />

also striking that only a rather small number of reports from<br />

Danish universities, Statistics Denmark, independent advisory<br />

boards etc. have dealt with or discussed expenditure needs – at<br />

least in a more continuous way. Exceptions are e.g. Bunzel et al.<br />

(1980), where three university professors vigorously criticised the<br />

statistical methods used in Finansieringsudvalget (1978). Another<br />

example is Dilling-Hansen et al. (1991), who made a contribution<br />

concerning economic conditions in small, perhaps economically<br />

distressed, rural local governments. This was part of the<br />

foundation for the organisation of economically disadvantaged (in<br />

Danish “Det Skæve Danmark”) municipalities, but this<br />

organisation has itself consciously abstained from analysing<br />

expenditure needs (see Det Skæve Danmark (2000)). Finally, Det<br />

Økonomiske Råd (The Danish Economic Council) (2002) made a<br />

study of the economy of local governments and also discussed<br />

equalisation and the expenditure needs system. However, the<br />

council did not bring forward new statistical analyses and did not<br />

question the existing set-up; it criticised the complexity of the<br />

system and noted the difficulties in the precise measurement of<br />

expenditure needs.<br />

Besides the contribution from the academic world, a range of<br />

individual municipalities, formal or informal organisations of local<br />

governments etc. has published studies of the equalisation system,<br />

often insightful but of course also usually seen from the point of<br />

view of exactly those municipalities. 26<br />

26<br />

See Finansieringsudvalget (2004, ch. 15) for a list of contributions. In 2000, the<br />

minister in office set up a special “Dialogforum” (Forum of Dialogue) on<br />

79


Finally, there are a few examples of specialist work being carried<br />

out by order of the Finance Committee; for example, Professor<br />

Milhøj (1998) contributed to the Finance Committee on the subject<br />

of statistical methods to be used for analysing expenditure needs.<br />

2.3.3. Evaluation<br />

The Danish model differs clearly from the Australian model, which<br />

builds on external, “objective” determinants of expenditure needs.<br />

It also differs from other countries in that the system of grants and<br />

equalisation is very much centralised internally in the government<br />

with MIT (supported by MOF) playing a strong role.<br />

How does this corporate model with strong coordination internally<br />

in government – both on technical matters and assessment of<br />

alternatives - work<br />

2.3.3.1. The corporate mode in Denmark<br />

There is in Denmark a general acceptance of a corporate model.<br />

Decisions are normally made following a hearing of the interested<br />

parties and negotiations. The Municipal association (KL) has been<br />

pushing this model cleverly since its creation in 1970 and has thus<br />

gained strong influence. The Finance Committee is one result of<br />

this policy. In general, however, close cooperation with the local<br />

governments‘ associations takes place in the so-called “budget<br />

cooperation system” (central/local consultation and budget<br />

cooperation system). The system – which involves frequent<br />

meetings between ministers and the chairs of the associations as<br />

well as annual negotiations following a more or less fixed schedule<br />

- has been functioning since the 1970’s, and today it is an<br />

important part of the Danish economic policy system. 27<br />

Why this cooperative (or even corporate) model has developed in<br />

Denmark is not clear, but part of the explanation is the fact that<br />

Denmark is a small and homogenous country with a certain degree<br />

of political consensus. Often members of the opposition in<br />

Parliament disagree with the agreements made in such forums,<br />

but know their usefulness if - and when - they get in office.<br />

equalisation, parallel with the Finance Committee. The Forum was, however,<br />

closed down already in 2001.<br />

27<br />

See e.g. Blom-Hansen (1998) and Lotz (2007).<br />

80


2.3.3.2. The Finance Committee and reliability<br />

No external evaluation of this aspect of the work of the Finance<br />

Committee has been carried out, but the lack of academic or<br />

alternative contributions may indicate that the committee has<br />

gained some degree of reliability. The frequent publication of<br />

reports may have contributed to the transparency of the work in<br />

this respect. 28<br />

Also, it is worth noting that in spite of Denmark most often being<br />

ruled by minority governments it has been possible to find a<br />

parliamentary majority for reforms of equalisation, building on the<br />

proposals from the Finance Committee with respect to expenditure<br />

needs. The opposition has been critical of the latest equalisation<br />

reform, which focuses on the final distributive outcome for<br />

individual local governments. It is not clear whether this also<br />

includes discontent with the needs criteria. One of the largest<br />

municipalities has expressed concerns about some of the new<br />

criteria, see Holdt-Olesen (2006).<br />

2.3.3.3. Future orientation or stifling of the system<br />

The fact that the expenditure needs system has changed, but only<br />

in intervals of 5-10 years, demonstrates some kind of stability. On<br />

the other hand – and just as importantly – the changes in some<br />

intervals indicate that the system is not stifled.<br />

It must also be stated, however, that the tasks of Danish local<br />

governments have become more sophisticated and the level of<br />

ambition concerning equality has risen (especially concerning the<br />

now solely block grant-financed regions); this may in the future be<br />

an important challenge to the ruling concept.<br />

2.3.3.4. The aspect of pragmatism<br />

The Danish model of committees consisting of civil servants seems<br />

to have given this aspect a high priority. Of course, this<br />

characteristic of the model, including a bias of a certain<br />

understandable ‘conservatism’ from the part of the organisations,<br />

28<br />

The report by Finansieringsudvalget (1998) was published as a kind of reaction<br />

to the 1995 reform, which was not prepared by the committee (see above), and did<br />

also raise criticism of certain needs criteria implemented after this reform, see<br />

op.cit. p.26.<br />

81


does on the other hand pose a threat of overlooking the other<br />

aspects. If e.g. pragmatism is favoured at the expense of the<br />

element of future orientation, this could be dangerous to the<br />

organisation in the longer run.<br />

If pragmatism is the headline, this could also jeopardise the<br />

reliability of the system if the public gets the impression that<br />

compromises, ‘horse trading’ and incrementalism stand in the way<br />

of sound, but more profound changes. However, here the<br />

participation of local government associations has undoubtedly<br />

been to some help in securing that ‘objective’ criteria are<br />

candidates for expenditure needs.<br />

2.4. Final comments<br />

The Danish system of equalisation of expenditure needs is seen as<br />

a relevant and integrated part of the general financial system. Due<br />

to the 2007 structural reform, larger local governments have<br />

levelled out some of the former differences, see the introduction.<br />

But with the recently implemented structural reform, the<br />

municipalities have instead received certain more sophisticated<br />

tasks from other levels of government, and the regions are now<br />

almost solely financed by block grants, which presents a challenge<br />

to the expenditure needs system.<br />

Over the years the way of thinking about expenditure needs has<br />

evolved. The normative production function approach has been<br />

ruled out in favour of the expenditure approach, aiming at<br />

measuring needs from observed data, correcting as well as possible<br />

for local income levels and preferences. However, being different<br />

from the normative approach, the expenditure method contains a<br />

problem of supply-side factors.<br />

Statistical analyses are used to identify needs criteria. They are,<br />

however, only one step of a longer procedure involving<br />

identification of potential needs criteria according to certain<br />

requirements of objectivity and choosing the best criteria taking<br />

several factors into account. The procedure secures among other<br />

things a stable system and – hopefully – a system which can be<br />

“translated” to others than experts in equalisation.<br />

82


Also the organisation of work has evolved. The technical work is<br />

centred on the Finance Committee and involves the Ministry of the<br />

Interior/the Ministry of <strong>Social</strong> Welfare, the Ministry of Finance, line<br />

ministries and local governments’ organisations, but only to a modest<br />

degree external, e.g. academic, contributions. The corporate model of<br />

organising the work of establishing proposals for new expenditure<br />

needs criteria via the Finance Committee is evaluated according to<br />

three parameters: reliability, future orientation and pragmatism. The<br />

ultimate test is, of course, the successful implementation of proposals<br />

from the committee.<br />

References<br />

Birch Sørensen, P., 1987: »Den kommunale indkomstbeskatning i Danmark«, memorandum<br />

til det af Indenrigsministeriet nedsatte udvalg vedrørende<br />

kommunernes finansiering. Københavns Universitet.<br />

Blom-Hansen, Jens, 1998: »Studier i statens styring af den kommunale økonomi«,<br />

Politicas PhD serie, Aarhus Universitet.<br />

Bunzel, H., S. Hylleberg & J. Søndergaaard, 1980: »Fordelingen af statens<br />

generelle tilskud«. Nordisk Administrativt Tidsskrift, nr.1.<br />

Council of Europe/Europarådet, 1992: »Equalisation of resources between local<br />

authorities«, Recommendation No. R(91) 4 and explanatory<br />

memorandum, Strasbourg.<br />

Det Skæve Danmark, 2000: »Oplæg til ændring af: Uddeling af statstilskud,<br />

udligning af beskatningsgrundlag, udligning af udgiftsbehov, udligning<br />

af købsmoms«.<br />

Det Økonomiske Råd, 2002: »Dansk Økonomi. Forår 2002«, rapport.<br />

Dilling-Hansen, M., K.R. Pedersen & V. Smith, 1991: »Det Skæve Danmark. En<br />

undersøgelse af økonomisk klemte kommuners indkomst-, erhvervs-,<br />

uddannelses- og serviceniveauer«.<br />

Finansieringsudvalget, 1978: »Kommunale udgiftsbehov«. Betænkning nr. 855.<br />

Rapport fra arbejdsgruooe vedrørende bloktilskud og andre<br />

udligningsordninger.<br />

Finansieringsudvalget, 1982: »Kommunal udligning og tilskud til kommuner«,<br />

Betænkning 963. Redegørelse fra det af indenrigsministeriet nedsatte<br />

udvalg vedr. kommunernes finansiering.<br />

Finansieringsudvalget, 1985: »Konjunkturregulering og visse udligningstekniske<br />

problemer«, betænkning 1033.<br />

Finansieringsudvalget, 1993: »Mulige ændringer i den amtskommunale udligning<br />

af udgiftsbehov«, January 1993.<br />

Finansieringsudvalget, 1998: »Betænkning om kommunernes udgiftsbehov.<br />

Redegørelse fra arbejdsgruppe under Indenrigsministeriets<br />

Finansieringsudvalg«, betænkning nr. 1361.<br />

Finansieringsudvalget, 2004: »Et nyt udligningssystem«, Betænkning 1437, bind I<br />

og II. Indenrigs- og Sundhedsministeriets Finansieringsudvalg.<br />

83


Finansieringsudvalget, 2005: »Finansieringsudvalgets rapport om et nyt tilskudsog<br />

udligningssystem«, Indenrigs- og Sundhedsministeriets<br />

Finansieringsudvalg.<br />

Holdt-Olesen, Peter, 2006: “Det politiske spil om udligningsreformen”,<br />

Administrativ Debat nr. 1, april.<br />

Indenrigsministeriet, 1993: »Redegørelse om aktivitetsbestemte tilskud,<br />

forenkling af udligningssystemet, statstilskuddets fordeling«.<br />

Betænkning nr. 1250.<br />

Indenrigsministeriet, 1995: »Forslag til lov om ændring af lov om kommunal<br />

udligning og generelle tilskud til kommuner og amtskommuner«, L245.<br />

Indenrigsministeriet, 2000: »Størrelseseffekter i den kommunale sektor«. Maj<br />

2000.<br />

Indenrigsministeriet, 2000: »Kommunale nøgletal 2000 – mål for serviceniveau.<br />

Et debatoplæg«.<br />

Kabelmann Thomas & Niels Jørgen Mau Pedersen, 1999: »<strong>Measuring</strong> social<br />

expenditure needs – is it possible to be objective « , paper for IIPFconference,<br />

Moscow, 1999.<br />

Lotz, Jørgen, 1987: »Kommunernes finansiering«, 3.udgave. Jurist- og<br />

Økonomforbundets forlag.<br />

Lotz, Jørgen, 2007: »Spillet om kommunernes økonomi«, in S. Lundtorp & M.<br />

Rasmussen (eds.): »Ledelse i kommunerne – på godt og ondt«.<br />

Handelshøjskolens Forlag. København.<br />

Mau Pedersen, Niels Jørgen & Inge Lene Møller, 2001: »Economics of scale in<br />

local governments – theoretical and empirical investigations of Danish<br />

municipalities«. Paper for the 57 th Congress of International Institute of<br />

Public Finance in Linz, Austria, August 2001 (mau@vfm.dk).<br />

Mau Pedersen, Niels Jørgen, 2007: »Den offentlige sektor i flere niveauer«,<br />

Jurist- og Økonomforbundets forlag.<br />

Milhøj, Anders, 1998: »Betænkning om kommunernes udgiftsbehov. Bilag om<br />

metodediskussion. « Indenrigsministeriet.<br />

PLS-Rambøll, 2000: »Regionaløkonomisk analyse af etableringen af Nakskov<br />

Industri- og Miljøpark«.<br />

Sundhedsministeriet, 1992: »Københavns og Frederiksberg kommuners<br />

sygehusudgifter«, June 1992.<br />

Spacojevic, John, 2008: »Fiscal equalisation in Australia – some technical issues«.<br />

Boerboom, Hessel & Peter Huigsloot, 2008: »The equalised allocation of local<br />

expenditure needs in the Netherlands – an optimalized mixture of<br />

objectivity and politics«.<br />

84


Appendix: Weights of expenditure needs of Denmark, Norway<br />

and the Netherlands<br />

Table. <strong>Expenditure</strong> needs criteria in equalisation systems in<br />

Denmark, Norway and the Netherlands, 2007<br />

Denmark<br />

DK Norway<br />

Norway Netherlands NL<br />

<strong>Expenditure</strong> Weight <strong>Expenditure</strong> Weight <strong>Expenditure</strong> Weight<br />

needs criteria per<br />

cent<br />

needs criteria per<br />

cent<br />

needs criteria per cent<br />

Age groups:<br />

0-6<br />

7-16<br />

17-19<br />

20-24<br />

25-34<br />

35-39<br />

40-64<br />

65-74<br />

75-84<br />

85 years or more<br />

Commuting time<br />

9.8<br />

21.2<br />

1.1<br />

2.0<br />

5.4<br />

2.9<br />

11.7<br />

4.2<br />

5.6<br />

4.8<br />

1.4<br />

Basic subsidy<br />

Age groups<br />

0-15<br />

6-15<br />

16-66<br />

67-79<br />

80-89<br />

90 years or more<br />

16-59 years<br />

divorced<br />

16-59 years<br />

unemployed<br />

2.5<br />

2.3<br />

30.8<br />

12.0<br />

8.5<br />

13.3<br />

4.9<br />

3.8<br />

1.1<br />

1.5<br />

Inhabitants<br />

Dwellings<br />

Age group 0-<br />

19 years<br />

<strong>Local</strong> regional<br />

functions<br />

High density<br />

Low income<br />

individuals<br />

23<br />

14<br />

10<br />

9<br />

9<br />

5<br />

20-59 years<br />

unemployed ><br />

5%*<br />

24-49 without<br />

vocational<br />

training*<br />

Rented dwelling*<br />

Psychiatric<br />

patients*<br />

Families in<br />

certain types of<br />

dwellings*<br />

Children of poorly<br />

educated<br />

parents*<br />

Singles more<br />

than 65 years<br />

5.4<br />

5.2<br />

1.5<br />

1.5<br />

4.5<br />

4.5<br />

0.7<br />

Commuting time<br />

Travel distance I<br />

Travel distance<br />

II<br />

Mortality<br />

Single 67 years<br />

or more<br />

Immigrants<br />

16 years or more<br />

mentally<br />

handicapped<br />

Less than 16<br />

years mentally<br />

handicapped<br />

1.0<br />

1.1<br />

2.5<br />

2.5<br />

0.5<br />

6.6<br />

0.4<br />

4.2<br />

0.5<br />

<strong>Social</strong> cash<br />

benefits<br />

<strong>Social</strong> support<br />

Minorities<br />

Regional<br />

regional<br />

functions<br />

Pupils<br />

secondary<br />

education<br />

Selected<br />

physical<br />

features<br />

Population<br />

characteristics<br />

Tax capacity<br />

5<br />

5<br />

4<br />

3<br />

3<br />

17<br />

7<br />

-20<br />

85


old*<br />

Low income<br />

individuals*<br />

Handicapped*<br />

Immigrants and<br />

descendants*<br />

Number of living<br />

years lost*<br />

3.0<br />

1.5<br />

0.7<br />

0.7<br />

0.7<br />

Degree of<br />

urbanization<br />

Rural districts<br />

Others 6<br />

Decline in<br />

populations<br />

numbers*<br />

Total 100.0 100.0 100.0<br />

Sources: Denmark: Velfærdsministeriet, Norway: Kommunal- og<br />

Regionaldepartementet, The Netherlands: Ministerie van Binnenlandse Zaken.<br />

86


Chapter 3<br />

The Use of <strong>Expenditure</strong> <strong>Needs</strong>: Equalization<br />

or Regional Policies<br />

Junghun Kim<br />

3.1. Introduction<br />

The size of local expenditures in Korea is larger than the central<br />

government’s expenditure. At the same time, many local<br />

governments, especially small ones, rely heavily on<br />

intergovernmental transfers. Even large cities, except those in the<br />

Seoul capital region, receive a large amount of intergovernmental<br />

transfers from the central government. The heavy financial<br />

dependence of local governments on the central government has<br />

been a long-time characteristic of the intergovernmental fiscal<br />

relations in Korea, but it has been an increasing trend since the<br />

start of decentralisation in 1995.<br />

It is recognised in the fiscal federalism literature that too much<br />

reliance on intergovernmental transfers can make the local public<br />

sector inefficiently too large as well as create adverse incentives<br />

(e.g. Oates, 2006). It is therefore an important policy issue to make<br />

the system of intergovernmental transfers as efficient as possible.<br />

Among many types of intergovernmental transfers in Korea,<br />

general-purpose grants called <strong>Local</strong> Allocation Tax (LAT) are made<br />

for equalisation across local governments, and are distributed<br />

based on the difference between “basic fiscal revenue” and “basic<br />

expenditure needs.” However, in some countries such as Canada,<br />

expenditure needs are not taken into account in the determination<br />

of equalisation grants. The argument for such an approach,<br />

supported by e.g. Shah (2006) and Boadway (2006), is that<br />

subsidising local governments with high cost structure is<br />

inefficient, and defining “basic needs” involves ambiguous<br />

87


concepts. This criticism does not imply that the central government<br />

should not support local governments’ provision of basic public<br />

services such as health and education. It rather implies that<br />

complicated methods to take into account local-specific cost<br />

structures can obscure policy objectives due to lack of transparency<br />

and simplicity.<br />

In this respect Korea has much room for improvement. Between<br />

the two components that constitute the general purpose grants,<br />

basic expenditure needs dominate the distribution of the grants<br />

with its weight being about 65 per cent. Also, the calculation of<br />

basic expenditure needs is very complicated, involving seventeen<br />

expenditure categories as well as complicated modification<br />

coefficients. The reason why expenditure needs are important in<br />

the system of general purpose grants is that it is used to ensure a<br />

“national minimum” for all local governments. However, the<br />

definition of a national minimum is not very clear, and the main<br />

effect of supporting the national minimum is to concentrate the<br />

grants towards the lagging regions. In this sense, the objective of<br />

regional policy as well as equalisation is embedded in the grants<br />

system in Korea. Therefore, if the two policy goals are separated,<br />

intergovernmental fiscal relations will become simpler and more<br />

transparent.<br />

The paper is organised as follows. Section 2 overviews the local<br />

public finance in Korea. The structure and characteristics of<br />

equalisation grants are examined in section 3. In section 4, the<br />

regional policy element embedded in equalisation grants is<br />

discussed, and section 5 concludes the paper.<br />

3. 2. Overview of local public finance<br />

3.2.1. <strong>Government</strong> structure<br />

Korea is a unitary country with 16 prefectures and 230<br />

municipalities. The process of decentralisation started in 1995, and<br />

the heads of the local governments are elected by popular vote<br />

every four years. Prefectures are composed of seven major cities<br />

88


and nine provinces. 29 Municipalities consist of special districts<br />

within the major cities, and the cities and towns within the<br />

provinces. There are currently 75 cities and 86 towns in the nine<br />

provinces, and 25 and 44 special districts in Seoul and other major<br />

cities, respectively.<br />

3.2.2. Population distribution<br />

The population sizes of the local governments are very diverse,<br />

with those of Seoul and Gyeonggi being around ten million, and<br />

that of Jeju, an island province south to the mainland of Korea,<br />

half a million. But the most notable aspect of population<br />

distribution in Korea is the population concentration in the capital<br />

region. Including Incheon, the fourth largest city 30 kilometres<br />

west of Seoul, almost 50 per cent of Korea’s population live in the<br />

capital region. The trend of population concentration has been<br />

going on for the past forty years. In the 1970s, the population<br />

share of the capital region was less than 30 per cent, but it<br />

increased to more than 35 per cent in the 1980s, and then up to 45<br />

per cent by the mid- 1990s.<br />

When we analyse the structure of local expenditure needs in<br />

Korea, it is important to recognise this rapid urbanisation process,<br />

since it is related to the strong element of the regional policy in the<br />

system of intergovernmental grants. Figure 1 shows the population<br />

share of local governments by ascending order of their population<br />

sizes. In 1990, about 30 per cent of the total population lived in 168<br />

relatively small local governments (80 per cent of the then total of<br />

210 municipalities). In 2005, about 25 per cent of the total<br />

population lived in 137 relatively small local governments (80 per<br />

cent of the total of 172 municipalities). Despite this rural-urban<br />

migration process, the share of general-purpose grants for the<br />

relatively small municipalities has even increased in the past<br />

decade. 30<br />

The reason for this is not just the central government’s concern<br />

about the welfare of people in the rural areas: the central<br />

29 Seven major cities are Busan, Daegu, Incheon, Gwangju, Daejeon, and Ulsan.<br />

Nine provinces are Gangwon, Gyeonggi, Chungbuk, Chungnam, Jeonbuk,<br />

Jeonnam, Gyeongbuk, Gyeongnam, and Jeju.<br />

30<br />

A more detailed discussion follows in Section 4.<br />

89


government is also concerned about rural-urban migration,<br />

especially the migration to the capital region. The central<br />

government has been engaged in the so-called “capital region<br />

policy” that tries to suppress migration to the capital region since<br />

the early 1970s. Such a regional policy is still very important in<br />

Korea, and it has been embedded in the structure of<br />

intergovernmental grants for a long time. 31<br />

Figure 1. The population share of local governments by decile<br />

3.2.3. <strong>Local</strong> revenue and intergovernmental grants<br />

According to the local budget report published by the Ministry of<br />

Public Administration and Security (MOPAS), the total revenue of<br />

local governments in 2007 was 112 trillion Won. 32 On average,<br />

65.9 per cent of the local revenue comes from own revenue sources,<br />

and 42.6 per cent from intergovernmental grants. 33<br />

31<br />

The previous President No Mu-Hyun seized power in 2003 with a campaign<br />

proposal of establishing a new capital town about 200 kilometers south of Seoul.<br />

The proposal led to an act, and all the administrative bodies are scheduled to<br />

move to the new capital town in early 2010s.<br />

32<br />

The exchange rate of one dollar is currently about 1,000 Won and therefore it is<br />

approximately $112 billion.<br />

33<br />

<strong>Local</strong> debt issuance is about 3.5 per cent of the total local revenue.<br />

90


However, the degree of fiscal dependency is diverse among local<br />

governments, and it is much higher than average for many small<br />

local governments. 34<br />

Currently, there are five types of intergovernmental grants in<br />

Korea. The most important are general-purpose grants, managed<br />

by two ministries: MOPAS manages the so-called <strong>Local</strong> Allocation<br />

Tax (LAT) to support the “basic needs” of the local administration, and the<br />

Ministry of Education and Science Technology (MEST) manages the<br />

<strong>Local</strong> Education Allocation Tax (LEAT). 35 The sizes of these<br />

grants are determined by law, and it is 19.24 per cent of the<br />

Domestic Tax Revenue for LAT, and 20 per cent for LEAT. 36 The<br />

share of Domestic Tax Revenue in the national tax revenue<br />

fluctuates depending on economic conditions and changes in the<br />

tax system, but it has been around 80 most of the time. Therefore<br />

about 32 per cent of the national tax revenue is distributed to the<br />

local public sector in the form of general-purpose grants.<br />

Another important item in the system of intergovernmental grants<br />

in Korea is specific grants managed by line ministries. There was a<br />

major reform in 2005 to reduce the size of specific grants. As a<br />

result, two types of new intergovernmental grants were created:<br />

the specific grants that were thought to be closely related to<br />

regional development projects became Balanced Development<br />

Special Account (BDSA); the specific grants that were related to<br />

subsidising the facilities for social welfare became so-called<br />

Decentralization Grants (DG). After this reform, the size of specific<br />

grants was reduced by about 37 per cent: the sizes of BDSA and<br />

DG as the shares of the pre-reform specific grants were 29 per cent<br />

and 8 per cent, respectively. 37<br />

34<br />

The share of cities' own revenue is 39.5% and that of towns 16.6%. On the other<br />

hand, the share of Seoul's own revenue is about 90 per cent and that of major<br />

cities about 62 per cent.<br />

35<br />

Education services are provided by education offices established and directly<br />

controlled by MEST.<br />

36<br />

Domestic Tax Revenue in Korean tax system is defined as the national tax<br />

revenue minus earmarked taxes such as gasoline tax, Education Tax, liquor tax,<br />

etc.<br />

37<br />

Most of BDSA and DG grants are distributed to local governments based on<br />

formulae.<br />

91


There is another type of intergovernmental grants, which was<br />

created in 2005 as a result of property tax reform. Unlike the case<br />

in other countries, property taxes in Korea are highly<br />

progressive. 38 In the 2005 reform, the property tax base became<br />

the assessed market value of the properties, and the number of the<br />

tax brackets was reduced to six. At the same time, the central<br />

government started to levy its own property tax on residential<br />

properties, applying the upper three brackets of the new property<br />

tax. 39 The revenue of the national property tax, called<br />

Comprehensive Real Estate Property Tax (CREPT), is distributed<br />

to local governments under the name of Property Tax<br />

Intergovernmental Grants (PTIG). The size of PTIG, being about 2<br />

billion won as of 2007, is not very significant compared to LAT and<br />

the specific grants, but it has been increasing quite fast for the<br />

past three years because of the national government’s effort to<br />

increase the revenue from property taxes.<br />

3. 3. Structure of equalisation Grants<br />

3.3.1. Formula of LAT<br />

The calculation of LAT involves three steps. Firstly, the total<br />

amount is set at a fixed percentage of Domestic Tax Revenue; this<br />

is currently 19.24 per cent. 40 Secondly, the fiscal gap is calculated<br />

based upon the difference between Basic Fiscal Revenue (BFR) and<br />

Basic <strong>Expenditure</strong> <strong>Needs</strong> (BEN). Thirdly, since the sum of the<br />

differences between BFR and BEN is usually greater than the<br />

predetermined size of LAT, the difference is scaled down by<br />

multiplying an adjustment factor, which is the ratio of the sum of<br />

the differences to the predetermined total amount of LAT. 41<br />

38<br />

Until 2004, the property tax on land had nine tax brackets with tax rates<br />

ranging from 0.2 per cent to 5 per cent; property tax on the structure of<br />

residential property had six tax brackets with the tax rates ranging from 0.3 per<br />

cent to 7 per cent.<br />

39<br />

That is, the property tax rate applied by local governments becomes flat above a<br />

certain threshold, but the central government applies progressive tax rates<br />

thereafter.<br />

40<br />

The ratio has increased three times since 2000. In 2001, it increased from 13.27<br />

per cent to 15 per cent. It rose to 19.12 per cent in 2005, and then to 19.24 per<br />

cent in 2006.<br />

41<br />

General purpose grants for educational services are provided to education<br />

offices, not local governments. Also, its distribution is almost on a per capita<br />

92


Basic Fiscal Revenue<br />

The calculation of BFR is relatively straightforward and is set at<br />

80 per cent of the local tax revenue. Loosely speaking, it is based<br />

upon the concept of Representative Tax System, since article 8 of<br />

the <strong>Local</strong> Allocation Tax Act stipulates that “standard local tax<br />

rates” are applied to calculate the BFR. “Standard local tax rate”<br />

refers to the local tax rate stipulated in the <strong>Local</strong> Tax Act, which is<br />

passed in parliament.<br />

One distinct aspect of local public finance in Korea is the fact that<br />

local tax rates are almost uniform across the country, and<br />

parliament is often blamed for this, as it sets the standard local tax<br />

rates. However, it should be noted that the <strong>Local</strong> Tax Act allows<br />

local governments to change, by their own by-laws, the local tax<br />

rates within certain boundaries. Despite the taxing power thus<br />

granted to local governments, albeit within a certain range, almost<br />

no local governments exercise it. Therefore, the calculation of BFR<br />

is easily done since it is very close to the actual local tax revenue<br />

raised. 42<br />

Although local tax rates are almost uniform across local<br />

governments, the per capita local tax revenue varies between them<br />

since many local tax items, including a surtax on corporate income<br />

tax, are related to business activities. Business activities, in turn,<br />

depend on agglomeration economies, and as a result, the size of the<br />

population is the main indicator reflecting the economic strength<br />

and the fiscal capacity of local governments in Korea. In Figure 4,<br />

where the relationship between the per capita local tax revenue<br />

and the population size of 16 prefectures is shown, it can be seen<br />

that the per capita tax revenues of Seoul and Gyeonggi are almost<br />

twice the average local tax revenue. This is because these regions<br />

basis. So we focus our discussion on the equalisation grants administered by<br />

MOPAS<br />

42<br />

Although this is not the main topic of the paper, the reason why local<br />

governments do not exercise taxing power is an important subject that requires<br />

in-depth analysis. In principle, one can hypothesise that they are related both to<br />

the adverse incentives caused by the system of intergovernmental grants and to<br />

yardstick competition among local governments.<br />

93


enjoy agglomeration economies and attract labour and capital from<br />

other parts of the country. 43<br />

Basic <strong>Expenditure</strong> <strong>Needs</strong><br />

The calculation of BEN involves 17 expenditure categories. For<br />

each expenditure category, three components are used for<br />

calculating the BEN: workloads ( ) such as the number of<br />

population and local officials, unit costs ( ), and modification<br />

factors( ). Thus BEN for expenditure category for local<br />

government can be expressed as the following (superscript<br />

denotes local government and subscript expenditure category):<br />

Figure 2. Per capita local tax revenue (Million Won)<br />

A pronounced feature of local expenditure needs in Korea is the<br />

emphasis on economies of scale in its calculation. In Figure 3, the<br />

distributions of the three components of the BEN of “local<br />

administration”, which is the largest expenditure category, are<br />

illustrated as a function of the population size: Figure 3-a shows<br />

the distribution of the per capita “standard” number of local<br />

43<br />

Ulsan is the city where large factories such as Hyundai car and ship-building<br />

companies are located. So its per capita local tax revenue is higher than that of<br />

the rest of the major cities.<br />

94


officials, 44 Figure 3-b the unit costs that are set at two different<br />

levels; and Figure 3-c the modification factors. As is evident from<br />

these figures, a very strong assumption of economies of scale is<br />

adopted in the calculation of BEN, and the final result is a strong<br />

inverse relationship between per capita BEN for local<br />

administration and the population size (Figure 3-d).<br />

Figure 3. Distributions of the components of BEN<br />

It is worth emphasising that the economies of scale effect is an<br />

assumption in the calculation of BEN. LAT has been in operation<br />

for the past 30 years, including the period of centralisation. Before<br />

1995, the year when local autonomy started, local governments’<br />

expenditure patterns were determined by the central government.<br />

This means that local governments’ independent decision-making<br />

44<br />

The “standard” number of local officials is set by the central government to<br />

avoid the moral hazard problem associated with local governments’ incentive to<br />

increase the number of local officials. However, the fact that equalisation grants<br />

increase with the number of local officials is the legacy of the long era of<br />

centralisation in Korea.<br />

95


on local expenditure was absent until then, and therefore the<br />

economies of scale effect determined by the production technique<br />

available in the local market was not known to the central<br />

government. Since 1995, the system of intergovernmental fiscal<br />

relationship has barely changed. Therefore, the modification<br />

factors presumed to reflect the economies of scale effect have long<br />

been embedded in the design of LAT, and its major role has been to<br />

financially support rural areas. In other words, modification<br />

factors are an important mechanism that transforms generalpurpose<br />

grants into regional policy grants.<br />

3.3.2. Statistical characteristics<br />

Since the impact of modification factors on the BEN is so strong, a<br />

regression of the per capita BEN on population size shows a very<br />

high degree of goodness of fit. Figure 4-a, which depicts the<br />

relationship between per capita BEN and population size in log<br />

scales, suggests that there is a cubic relationship between the per<br />

capita BEN and the population size. Therefore, if we run the<br />

regression of per capita BEN ( ) on population ( ), population<br />

squared ( ) and population cubed( ), we get the results in (1)<br />

(with standard errors in parentheses). Besides population size, the<br />

area of local government is another important variable that affects<br />

BEN. Therefore, if the area of local government (area) is added as<br />

an explanatory variable, we get the results in (2):<br />

(1)<br />

(2)<br />

All the coefficients in the regressions are significant at a one<br />

percent significance level. Also the values of R-squared are very<br />

high: the population variable alone explains as much as 92 per cent<br />

of the variations in the per capita BEN, and the explanatory power<br />

of the regression model increases to more than 97 per cent if an<br />

area variable is added. Figure 7-b shows the relationship between<br />

per capita BEN and population size in a raw scale along with the<br />

fitted values. It shows that the per capita BEN becomes almost a<br />

constant when the population size exceeds around 500,000.<br />

96


Figure 4. Distribution of BEN with regression fit<br />

3. 4. Discussions<br />

3.4.1. <strong>Needs</strong> equalisation or regional policy<br />

The economies of scale effect in the production of local public goods<br />

is an important factor in the calculation of local expenditure needs<br />

in other countries such as Switzerland and Japan. 45 Therefore, it<br />

cannot be a priori argued that it is a peculiar aspect found only in<br />

Korea. However, what is distinct in the case of Korea is that<br />

migration from rural to urban areas, which supposedly exploits the<br />

economies of scale effect, has not reduced the financial burden of<br />

the central government. In Figure 5, the shares of general purpose<br />

grants received in the years 1990 and 2005 by a group of local<br />

governments are shown in ascending order by their population<br />

sizes. As discussed previously, a sizable migration from smaller<br />

local governments to larger ones took place during the period.<br />

However, the share of general purpose grants provided to the<br />

smallest 10 per cent and 20 per cent of local governments is<br />

greater in 2005 than in 1990. Therefore, even though the<br />

population share of small local governments has significantly<br />

decreased since 1990 46 , the share of the general purpose grants<br />

they receive has somewhat increased. This fact becomes more<br />

obvious if viewed from another perspective: most of the local<br />

governments in the capital region do not receive LAT because their<br />

45<br />

See Dafflon (2006) and Rechovsky (2006).<br />

46<br />

Since 1990, almost 4 million have moved into capital region.<br />

97


BFRs are greater than BENs. This means that the amount of LAT<br />

required to support local governments outside the capital region<br />

becomes smaller to the extent that migration to the capital region<br />

is taking place. However, the size of equalization grants as a share<br />

of Domestic Tax Revenue has increased from 13.27 per cent to<br />

19.24 per cent since 1990. It is therefore obvious that the share of<br />

LAT for rural areas has increased despite extensive migration to<br />

the capital region.<br />

Figure 5. The share of equalisation grants by decile<br />

Since general-purpose grants in Korea are strongly influenced by<br />

the agglomeration of migrants in the capital region, an important<br />

question is whether regional policies that try to slow down<br />

migration to the capital region improve social welfare. The<br />

literature on this issue is comprehensive, and many studies find<br />

that such regional policies are not efficiency-enhancing. 47 Even on<br />

equity grounds, the effectiveness of regional policies are often<br />

questioned. 48<br />

Judging from these studies, the intergovernmental grants in<br />

Korea, which are all similar in terms of their distribution across<br />

local governments, do not seem to have a sound theoretical basis,<br />

except for the fact that they are politically demanded. Although<br />

this evaluation of intergovernmental grants in Korea can be<br />

47<br />

See, e.g., Martin (1999), Puga (2002) and Midelfart and Overman (2002).<br />

48<br />

See Dupont and Martin (2006) and Pfluger and Sudekum (2007) .<br />

98


generally true, there is another strand of literature that offers<br />

theoretical justification of the design of intergovernmental grants<br />

that mitigate interregional migration. According to Boadway and<br />

Flatters (1982) and Boadway (2004), interregional migration can<br />

make regional allocation of labour inefficient if it is induced by net<br />

fiscal benefit (NFB) differentials across local governments. 49 The<br />

case of NFB differentials, which is often discussed in the Canadian<br />

context, is rich resource endowments of some regions in Canada.<br />

Another possible cause of NFB differentials is the economies of<br />

scale effect in the consumption of local public goods. However, the<br />

latter does not have any practical importance in most OECD<br />

countries since the local governments in these countries provide<br />

public services such as education, health, social services, police,<br />

which local governments do not have the economies of scale in<br />

providing. 50<br />

The situation can be different in the case of Korea, however, since<br />

the typical local public services such as education and social<br />

services are not the responsibility of local governments. This<br />

means that expenditure responsibilities of a large local government<br />

such as Seoul does not increase proportionally with the size of<br />

population, whereas strong business activities in the large cities<br />

make the tax revenue increase more than proportionally with the<br />

size of population. Thus the gap between the fiscal benefit and<br />

fiscal burden created by the continuous migration to the capital<br />

region may be generating a sizable amount of net fiscal benefit for<br />

the large cities in Korea.<br />

If this is indeed the case, Korea can be characterised as having a<br />

very poor design of intergovernmental fiscal relations since it<br />

suffers from dual distortions: ill-designed tax assignment and<br />

expenditure assignment make the capital region a fiscally<br />

attractive region, and then the government spends a lot of tax<br />

49<br />

Suppose, g = GN - a<br />

where g = per capita services, G = local public spending,<br />

N = population in a local government, a = degree of privateness, and t = per<br />

capita local tax. If there are economies of scale in providing local public goods, or<br />

if a ® 0 , then g » G > t . See Boadway and Flatters (1982) and Boadway<br />

(2004) for detailed discussions on the concept of net fiscal benefits.<br />

50<br />

Thus, Boadway (2004, 2006), for example, assume away the economies of scale<br />

in the provision of local public goods when analysing fiscally-induced migration.<br />

99


evenue on mitigating labour and capital migration to the capital<br />

region by inefficient and perhaps ineffective regional policies.<br />

3.4.2. Difficulties of reforming LAT<br />

Despite the strong criticism that can be directed at the role of LAT,<br />

its change is not easy in reality. First of all, since its structure is<br />

related to the fiscal attractiveness of the capital region, reducing<br />

the support for rural areas needs to go hand in hand with<br />

increasing the fiscal burden of the large cities, especially the Seoul<br />

region, by making them more responsible for local public services.<br />

However, this will be an extremely difficult task because local<br />

governments in the capital region are politically strong, and at the<br />

same time the emotional support for rural areas is also strong.<br />

The second reason why the reform of LAT is difficult is more<br />

fundamental. The equalisation grants in Korea are called <strong>Local</strong><br />

Allocation Tax. It will be confusing to outside observers why grants<br />

are called a “tax” in Korea. 51 This is indeed confusing even among<br />

the Korean people who use this terminology, since the <strong>Local</strong><br />

Allocation Tax is not even tax sharing, but purely revenue<br />

sharing. 52 So the reason why equalisation grants in Korea is called<br />

a “tax” is symbolic and political: it carries the impression that the<br />

local governments that receive LAT have a legitimate right to it, as<br />

if it was their own tax revenue. This is indeed the interpretation<br />

given by MOPAS, the ministry of the central government which<br />

administers LAT. Because of the strong institutional and political<br />

characteristics embedded in LAT, it is very hard to change its<br />

distribution pattern based on a pure cost-effectiveness argument.<br />

However, even if we accept the argument that LAT plays a larger<br />

role than just expenditure needs equalisation, it still has much<br />

room for improvement. Since general-purpose grants are<br />

distributed according to the rationale of equalising expenditure<br />

needs, the regional policy element of it has not been clearly<br />

identified. As a matter of fact, the role of regional policy grants is<br />

supposed to be played by different intergovernmental grants such<br />

51<br />

The same is true in Japan, whose system Korea has apparently adopted.<br />

52<br />

McLure(2001) defines tax sharing as the one different levels of governments<br />

share the same tax base on an origin basis. So he regards the revenue allocation<br />

among German landers as revenue sharing rather than tax sharing.<br />

100


as the <strong>Local</strong> Transfer Fund before 2004, and BDSA after 2005.<br />

However, if we compare the distribution patterns of LAT and the<br />

<strong>Local</strong> Transfer Fund, the regional policy element of LAT is even<br />

greater than that of the <strong>Local</strong> Transfer Fund since the<br />

concentration of grants to under-developed and less populated area<br />

is more pronounced in LAT than in the <strong>Local</strong> Transfer Fund. 53<br />

Therefore, it was difficult to distinguish between the different roles<br />

of the two types of grant, and ultimately the <strong>Local</strong> Transfer Fund<br />

was absorbed into LAT in 2005. Actually, this recent episode of<br />

merging the <strong>Local</strong> Transfer Fund into LAT makes it obvious that<br />

LAT plays the role of both expenditure needs equalisation and<br />

regional policy.<br />

The suggestion that LAT needs to be separated into an expenditure<br />

needs equalisation component and a regional policy component is<br />

also related to the fact that the dominance of the expenditure<br />

needs equalisation component in fiscal equalisation scheme is not a<br />

sound policy, as is argued by Shah (2006) and Boadway (2006).<br />

Especially Shah (2006) argues that expenditure needs equalisation<br />

is unnecessary since calculating expenditure needs is usually<br />

complicated and easily becomes discretionary. Boadway (2006)<br />

emphasises the importance of fiscal equalisation, especially in the<br />

case of NFB differentials, but he notes the importance of<br />

differentiating between costs equalisation and needs equalisation.<br />

<strong>Needs</strong> equalisation can be necessary on equity grounds, but it is<br />

hard to define “basic” local expenditure needs. More importantly, if<br />

costs equalisation is embedded in the fiscal equalisation scheme, it<br />

indirectly supports an inefficient provision of local public services.<br />

In the case of Korea, the fact that strong elements of financial<br />

support of local officials in LAT continue to exist demonstrates that<br />

expenditure needs equalisation promotes the inefficient behaviour<br />

of local governments.<br />

3.4.3. Decomposition of general-purpose grants<br />

The assumption of economies of scale that dominates the<br />

calculation of general-purpose grants in Korea makes the<br />

distribution of the per capita LAT significantly different between<br />

urban and rural regions. However, its impact on LAT distribution<br />

53<br />

This is because the main function of <strong>Local</strong> Transfer Fund was to support local<br />

road construction, of which the fast-growing Gyeonggi is in the greatest need.<br />

101


among rural local governments is not as significant. Table 1 shows<br />

a comparison of fiscal revenue of eight rural provinces. 54 The first<br />

row shows the total fiscal resources of the provinces, which consist<br />

of local taxes, LAT and <strong>Local</strong> Transfer Fund. In the second row, the<br />

per capita revenue is shown, and it is the lowest for Gyong-sang at<br />

1.33 million won and highest for Gangwon (1.88 million won),<br />

followed by Jeonnam (1.78 million won). Let us assume, therefore,<br />

that the central government evenly distributes financial resources<br />

to each province with 1.33 million to each. This will make a deficit<br />

of 0.8 trillion won for Gangwon, and 0.89 trillion won for Jeonnam,<br />

representing, respectively, 29 per cent and 25 per cent of the total<br />

revenue of the two provinces (row e). These differences can be<br />

interpreted as mainly stemming from such cost factors as<br />

mountainous areas, islands, etc., apart from the economies of scale<br />

that dominate the distribution of LAT between rural and urban<br />

regions. Therefore the differences can be separated from the<br />

equalisation grants and transformed into regional policy grants.<br />

Table 1. The revenue structure of rural regions<br />

Gang-<br />

Won<br />

Chun<br />

g-Buk<br />

Chun<br />

g-<br />

Nam<br />

Jeo<br />

n-<br />

Buk<br />

Jeon-<br />

Nam<br />

Gyong<br />

-Buk<br />

Gyong<br />

-Nam<br />

Total Revenue(tnwon) (a) 2.86 2.23 3.19 2.65 3.53 4.06 4.18 0.93<br />

Per capita(ml won) (b) 1.88 1.5 1.63 1.39 1.78 1.51 1.33 1.67<br />

b-Min(b) (c) 0.55 0.17 0.3 0.06 0.45 0.18 0 0.34<br />

c×pop(tn won) (d) 0.8 0.2 0.6 0.1 0.89 0.49 0 0.19<br />

100×d/a (%) 29.4 11.2 18.5 4.1 25.2 12.7 0 20.4<br />

Sum(d)/Sum(a) 14.2%<br />

Jeju<br />

It needs to be noted that this simulation is more illustrative than<br />

immediately applicable since it requires fundamental changes in<br />

intergovernmental fiscal relations, including a rearrangement of<br />

local tax bases and a more important fiscal role of provinces rather<br />

than of cities and towns. However, it shows that an allocation of<br />

financial resources to rural regions can be greatly simplified since<br />

about 85 per cent of the current financial resources assigned to<br />

these regions can be distributed on an equal per capita basis. In<br />

54 The Gyeonggi province is excluded from the comparison since it is situated in the capital<br />

region.<br />

102


this way, a great portion of general-purpose grants needs not be<br />

tied to vaguely defined indexes of expenditure needs.<br />

A similar calculation can be made for the urban regions, shown in<br />

Table 2.A notable difference between Table 2 and Table 1 is that<br />

the per capita revenue of the urban regions is significantly lower<br />

than that of the rural regions. As discussed previously, this is due<br />

to the assumption of economies of scale in the allocation of LAT.<br />

Since this paper focuses on the way to simplify the allocation of<br />

general-purpose grants, rather than on changing the financial<br />

flows from the relatively rich regions (urban regions) to poorer<br />

regions (rural regions), the difference of per capita revenue<br />

between urban and rural regions is taken for granted. Thus, in<br />

Table 2, the minimum per capita revenue of 0.65 million won can<br />

be assigned to all urban regions and, to keep the current financial<br />

resources unchanged, additional financial resources can be<br />

awarded to Seoul, Gyeonggi, and Ulsan.<br />

Table 2. The revenue structure of urban regions<br />

Seoul Busa<br />

n<br />

Dae-<br />

Gu<br />

Incheon<br />

Gwan<br />

g-ju<br />

Daejoen<br />

Ulsan<br />

Gyeo<br />

ng-gi<br />

Total<br />

9.74 2.48 1.66 1.77 0.91 1.03 0.99 9.98<br />

Revenue(tnwon)<br />

(a)<br />

Per capita(ml won) 0.96 0.68 0.66 0.69 0.65 0.71 0.92 0.95<br />

(b)<br />

b-Min(b) (c) 0.31 0.03 0.01 0.04 0 0.06 0.27 0.3<br />

c×pop(tn won) (d) 3.15 0.11 0.03 0.1 0 0.09 0.29 3.14<br />

100×d/a (%) 32.3 4.4 1.8 5.7 0 8.7 29.3 31.4<br />

Sum(d)/Sum(a) 24.2%<br />

The results in Table 1 and Table 2 are therefore opposite in the<br />

sense that richer urban regions are financially favoured while<br />

poorer rural regions are financially favoured. This contrast<br />

actually demonstrates the dual distortions mentioned above: local<br />

governments in the capital region (Seoul and Gyeonggi) are<br />

favoured by tax exporting elements in the design of local taxes, but<br />

they do not assume the responsibilities of providing local public<br />

services such as education, police, etc. At the same time, a lot of<br />

financial resources are devoted to controlling migration to the<br />

capital regions. An improvement in the intergovernmental fiscal<br />

103


elations in Korea will therefore involve changes in two directions.<br />

First of all, it needs to be recognised that tax assignment and<br />

expenditure assignment are related to fiscally induced migration to<br />

the capital region. Secondly, the regional policy element embedded<br />

in the intergovernmental grants system needs to be separated from<br />

the equalisation grants system so that the intergovernmental fiscal<br />

relations can be made simpler and more transparent.<br />

3. 5. Conclusion<br />

This paper addressed the issue of calculating expenditure needs in<br />

the system of general purpose grants in Korea. The most notable<br />

feature of expenditure needs in Korea is an inverse relationship<br />

between per capita expenditure needs and the size of the<br />

population, which implies that the general purpose grants heavily<br />

favours local governments with low population density.<br />

Another important feature of intergovernmental fiscal relations in<br />

Korea is the fiscal benefit given implicitly to the capital region.<br />

Major infrastructures such as airport, railways, and subways have<br />

been constructed around the capital region for a long time. Also,<br />

the current designs of tax assignment and expenditure assignment<br />

are such that they give rise to sizable net fiscal benefits to the<br />

residents living in the capital region. It is therefore not surprising<br />

that general purpose grants are, albeit implicitly, related to the<br />

regional policy objective, which is to control the agglomeration of<br />

migrants around the capital region.<br />

This paper has shown that, once such a structure is recognized, a<br />

simpler and more transparent design of general purpose grants is<br />

possible. If the roles of fiscal equalisation and fiscal policy are<br />

separated in the system of intergovernmental grants in Korea,<br />

policy objectives can be more clearly defined, and the effectiveness<br />

of both fiscal equalisation and regional policy can be enhanced.<br />

References<br />

Boadway, R.(2004), ‘ The theory and practice of equalization’, CESifo Economic<br />

Studies 50, 211-254.<br />

104


Boadway, R.(2006), Grants in a federal economy: A conceptual perspective, in R.<br />

Boadway & A. Shah, eds, ‘Intergovernmental Fiscal Transfers: Principles and<br />

Practice’, The World Bank, Washington.<br />

Boadway, R. & Flatters, F.(1982), ‘Efficiency and equalization payments in a<br />

federal system of government: A synthesis and extension of recent results’,<br />

Journal of Canadian Economics 15, 613-33.<br />

Dafflon, B.(2006), Fiscal capacity equalization in horizontal fiscal equalization<br />

programs, in R. Boadway & A. Shah, eds, ‘Intergovernmental Fiscal Transfers:<br />

Principles and Practice’, The World Bank, Washington.<br />

Dupont, V. & Martin, P.(2006), ‘Subsidies to poor regions and inequalities: some<br />

unpleasant arithmetic’, Journal of Economic Geography 6, 223-240.<br />

McLure, C.(2001), ‘The tax assignment problem: ruminations on how theory and<br />

practice depend on history’, National Tax Journal 52, 339-364.<br />

Midelfart, K. & Overman, H.(2002), ‘Delocation and european integration: is<br />

structural spending justified’, Economic Policy 35, 322-359.<br />

Oates, W.(2006), ‘Toward a second-generation theory of fiscal federalism’,<br />

International Tax and Public Finance 12, 349-373.<br />

Pfluger, M. & Sudekum, J.(2008), ‘Integration, agglomeration and welfare’,<br />

Journal of Urban Economics 63, 544-566.<br />

Reschovsky, A.(2006), Compensating local governments for differences in<br />

expenditure needs in a horizontal fiscal equalization program, in R. Boadway<br />

& A. Shah, eds, ‘Intergovernmental Fiscal Transfers: Principles and Practice’,<br />

The World Bank, Washington.<br />

Shah, A.(2006), A practitioners guide to intergovernmental fiscal transfers, in R.<br />

Boadway & A. Shah, eds, ‘Intergovernmental Fiscal Transfers: Principles and<br />

Practice’, The World Bank, Washington.<br />

Ulltveit-Moe, K.(2007), ‘Regional policy design: An analysis of relocation,<br />

efficiency and equity’, European Economic Review 51, 1443-1467.<br />

105


Chapter 4<br />

The equalised allocation of local<br />

expenditure needs in the Netherlands: an<br />

optimised mixture of objectivity and politics<br />

Peter Huigsloot<br />

Summary<br />

An equalised allocation of local expenditure needs has to be an<br />

optimised mixture of objectivity and politics. In the Netherlands,<br />

the aim of equal (not uniform) levels of services combined with a<br />

large degree of municipal autonomy is expressed in the allocation<br />

of financial means from the central government to municipalities:<br />

as many general grants as possible, reducing the number of<br />

specific grants, specific grants should have a wide target and a<br />

minor role for (redistributed) own incomes of municipalities.<br />

The most important instrument for the allocation of general<br />

means, the Municipal Fund, is based on a system of objective<br />

criteria applied to each municipality. The indicators in the<br />

Municipal Fund are related to large differences in costs between<br />

municipalities caused by varying exogenous factors (features of<br />

social, physical and regional structure in combination with legal<br />

regulations).<br />

The indicators applied in the Municipal Fund are based on a new<br />

method: the Difference Analysis. This new method has been chosen<br />

because of the proven serious shortcomings of the use of<br />

statistic/econometric techniques for the selection and weighing of<br />

criteria. This paper describes the method of the Difference<br />

Analysis leading to a well-balanced mixture of an objective<br />

substantiation of local financial needs and political decisions.<br />

106


4.1. General features of governmental and financial relations in<br />

the Netherlands<br />

4.1.1. The Dutch municipalities: uniformity in competences;<br />

variety in costs<br />

In the Netherlands, there are 443 municipalities with a varying<br />

number of inhabitants spanning from 1000 to 750,000. Despite this<br />

difference in the number of inhabitants, all municipalities are –<br />

with the exception of the four largest cities – uniform in<br />

competences and have the same tasks to fulfil. These tasks are<br />

supposed to be realised with a considerable degree of municipal<br />

autonomy.<br />

Among the municipalities, there are large dissimilarities in the<br />

volume, composition and costs of services related to the<br />

municipalities’ own discretion in respect of the fulfilment of tasks<br />

in combination with objective differences in demand, based on<br />

physical, social and regional features (see figure 1).<br />

In the social structure, the composition of the population in cohorts<br />

of age, health and socio-economic status (income, employment) is<br />

determinant. In the physical structure, the features of an urban or<br />

rural structure (a thinly populated rural authority versus a<br />

densely populated city authority), the kind of development<br />

(dense/thin, old/recent) and the soil condition are determining<br />

factors. In the regional structure, the functions performed by<br />

larger municipalities outside the scope of their own municipal area<br />

are of particular importance. These functions are mainly related to<br />

policy areas like art, sports and entertainment.<br />

107


Figure 1. Diversity of features of social, physical and regional<br />

structure<br />

4.1.2. Equal level of services, autonomy and decentralisation<br />

Despite the dissimilarities based on social, physical and regional<br />

structures, the Netherlands have high ambitions regarding the<br />

equalisation of services. All municipalities, despite their varying<br />

social, physical and regional features, must possess the financial<br />

potential to realise an equalised level of services, with an emphasis<br />

on equality rather than uniformity.<br />

Equalised levels of services are supposed to be realised according<br />

to the municipal autonomy as set out below. Limiting conditions<br />

presented by the central government in regard to the composition<br />

of services and spending are in particular realised by means of<br />

legal instruments (laws, general rules of government) rather than<br />

by means of financial instruments.<br />

The autonomy of the municipalities has been increasing in the past<br />

few years due to the decentralisation of tasks, for which purpose<br />

financial means are provided by general means (municipal fund) or<br />

by specific ‘wide target’ grants. Examples are the decentralisation<br />

of school accommodation, benefits under the recent <strong>Social</strong> Support<br />

108


Act (benefits for domestic care, for the disabled) and employment<br />

and income benefits.<br />

4.1.3. The financial allocation with a pivotal role for the municipal<br />

fund<br />

General grants, specific grants and autonomous income of<br />

municipalities<br />

The aim of equal levels of services in combination with a large<br />

degree of municipal autonomy is also expressed in the allocation of<br />

financial means from central to local government 55 : as many<br />

general grants as possible, a reduction in the number of specific<br />

grants, and giving specific grants a wide target. The autonomy of<br />

the Dutch municipalities as regards taxes is limited.<br />

Municipal fund and general grants<br />

The most important instrument for the allocation of general<br />

means, the municipal fund, is based on a system of objective<br />

criteria applied to each municipality. Only the four largest cities<br />

receive supplementary grants because of their exclusive tasks as<br />

large cities.<br />

Specific grants<br />

Apart from general means, municipalities also receive specific<br />

grants. In the past few years, the purpose of the policy has been to<br />

reduce the number of specific grants or to cluster the existing<br />

specific grants as much as possible into ‘wide target’ grants.<br />

Important specific grants are the benefits for employment and<br />

income (social security), urban renewal and the problems of large<br />

cities (in particular for public security), benefits for youth and<br />

family and for youth healthcare.<br />

In the last decades, the number of specific grants has decreased<br />

significantly: starting with more than 500 in 1980 to less than 200<br />

at present. The aim is to decrease the number even further to just<br />

a few dozen. The reduction of the number of specific grants did not,<br />

incidentally, go hand in hand with a similar reduction in the<br />

amount of the corresponding allocation. Most of the time, existing<br />

smaller specific grants were clustered into wider ones. The amount<br />

of allocated specific grants actually did not decrease very much.<br />

55. The same applies to the provinces<br />

109


Meanwhile a large part of the means from specific grants is not<br />

directly provided to municipalities but to regional authorities<br />

(including general law areas). Furthermore, there are differences<br />

in the significance of specific grants for individual municipalities.<br />

Some of the specific grants are only given to some municipalities,<br />

as regards urban renewal for instance.<br />

Similar criteria (indicators) may be used for the allocation of<br />

specific wide-target grants as for the general grants within the<br />

municipal fund. The range of spending possibilities is restricted,<br />

and municipalities are obliged to justify their spending to the<br />

central government.<br />

Autonomous municipal incomes<br />

The fact that municipalities have their own specific resources is<br />

taken into account when allocating general grants through the<br />

municipal fund.<br />

Municipalities can only generate their own income from property<br />

taxes by applying a higher rate than the central standard on which<br />

the municipal fund is based. The municipal fund corrects<br />

dissimilarities between municipalities in valuing property by<br />

means of a negative allocation formula. This negative allocation<br />

formula varies between municipalities from 6% to almost 30% of<br />

the general means.<br />

Moreover, municipalities are supposed to be able to cover<br />

approximately 4% of their general spending by means of other<br />

independent resources, including, in particular, revenues of<br />

investments, participations or land exploitations.<br />

Besides these independent general resources, municipalities are<br />

also authorised to charge cost-effective contributions to civilians<br />

(administrative charges, taxes or duties, etc.) on account of the<br />

provision of specific services. Examples of such contributions are<br />

waste collection levies, sewerage charges, charges for services<br />

provided by the population register and the granting of building<br />

and other permits.<br />

110


Relative limited volume of general grants, specific grants and<br />

independent incomes<br />

At present, the volume of the allocation of general means is<br />

approximately similar to the volume of the allocation of specific<br />

means (approximately 17 billion euro). The municipal own incomes<br />

(in particular from taxes, investments and charges) amount to<br />

approximately 30% of the general municipal means. There is a<br />

large variety in the volume of the allocated means between<br />

municipalities, both in regard to general means (see the next<br />

chapters) and to specific grants.<br />

4.2. Compartments and clusters in the Municipal Fund: volumes<br />

and features of allocation<br />

4.2.1. Survey of municipal tasks: 14 clusters of expenditure in four<br />

compartments<br />

The municipal fund distinguishes between 14 clusters of<br />

expenditure, subdivided into four compartments.<br />

The compartments and clusters have been logically organised (see<br />

figure 2):<br />

o the public area compartment includes the clusters of public<br />

parks and gardens, roads, water and sewerage;<br />

o the buildings and environment compartment includes the<br />

clusters of public order and security, museums etc., public<br />

housing, spatial planning and urban renewal, physical<br />

environment and garbage collection;<br />

o the public services compartment includes the clusters<br />

education, employment and income, social care (including for<br />

the disabled and children), art, sports and entertainment and<br />

civil registry;<br />

o the local government cluster.<br />

Relation with the condition of the soil, physical, social and regional<br />

structure<br />

The nature and intensity of tasks within the clusters vary between<br />

municipalities as a consequence of the effect of the features of<br />

social, physical and regional structures. Depending on their<br />

composition, the tasks and costs within a cluster depend on<br />

features of the physical structure (i.e. the condition of the soil, low<br />

or high density and quality of the buildings), features of the social<br />

111


structure(i.e. the composition of the population) and features of the<br />

regional structure (i.e. services in particular for art and<br />

entertainment used by citizens from outside the municipality).<br />

Figure 2. Compartments and clusters of the Municipal Fund and<br />

their relation with features of social, regional and physical<br />

structure<br />

112


4.2.2. Survey of means and criteria of allocation by compartment<br />

The Municipal Fund, with its system of indicators for the<br />

allocation of the majority of municipal financial means (the general<br />

means), is structured in accordance with the 14 clusters mentioned<br />

above. Each cluster has a formula of allocation, made up of one or<br />

more objective criteria (indicators). Figure 3 gives an overview of<br />

the individual clusters and the criteria of allocation that have been<br />

applied.<br />

Figure 3. Compartments, clusters and indicators of the Municipal<br />

Fund<br />

Figure 3 shows that different kinds of criteria are applied within<br />

the compartments.<br />

Public area<br />

The public area compartment shows physical criteria, including<br />

condition of the soil, surface area of land, surface area of buildings,<br />

volume of internal and external waterways, length of the shoreline<br />

and number of isolated areas within municipalities. These criteria<br />

are applied to the allocation of means for the clusters roads and<br />

113


water (in particular) and sewerage (just the condition of the soil). 56<br />

The means for public parks and gardens are not allocated on the<br />

basis of the latter criteria, but on the criteria of inhabitants and<br />

housing (see also chapter 5).<br />

Buildings and environment<br />

In regard to the allocation of means within the buildings and<br />

environment compartment and their distinguished imbedded<br />

clusters, the following criteria have been applied: number of<br />

homes, density of the building structure within a specific area (the<br />

number of addresses), number of business establishments, code for<br />

allocation of means for the benefit of urban renewal and several<br />

criteria related to the historical features of municipalities:<br />

inhabited regions 1931 (museums) and historical centres and<br />

waterways.<br />

<strong>Social</strong> services<br />

Within the social services compartment, the following criteria have<br />

been applied: number of inhabitants subdivided into cohorts of age,<br />

features of social structure (low incomes, persons on social<br />

security, persons drawing benefit(s), minorities, single-parent<br />

households), number of pupils in specific and secondary education,<br />

and criteria of central functions like art, sports and entertainment<br />

(i.e. local and regional customer potential).<br />

<strong>Government</strong><br />

Relevant criteria for the allocation of means for the local<br />

government are the criteria of inhabitants and the fixed amount. 57<br />

Fixed amounts and scale effects<br />

The fixed amounts are related to different compartments and are<br />

meant to compensate for the various scale effects of the smallest<br />

and the largest municipalities:<br />

o the smallest municipalities are compensated for negative scale<br />

effects (fixed costs, indivisibilities) of setting up the municipal<br />

administrative organisation;<br />

56<br />

.The major part of the means for sewerage charges is supposed to be paid from<br />

the proceeds of sewerage charges<br />

57.<br />

Exclusive criteria are used for the (West)Frisian islands.<br />

114


o<br />

the largest municipalities are paid for typical major city and<br />

provincial tasks that are exclusively carried out by the four<br />

largest cities.<br />

Relation with the VAT Compensation Fund<br />

Within the municipal fund it is taken into account that<br />

municipalities can claim the taxes they paid for governmental<br />

tasks from the so-called VAT compensation fund. The compensated<br />

amount is approximately 8% of their general means. The VAT<br />

compensation fund was created to stimulate the outsourcing of<br />

work. In the past, outsourced work was subject to VAT, whereas<br />

tasks carried out by the municipality itself were not or only to a<br />

very limited degree. As a result of the ability to claim VAT, both<br />

options are now in the same financial position.<br />

4.2.3. Results of the allocation of general grants to different kinds<br />

of municipalities<br />

Because the municipal fund recognises the differences in costs<br />

caused by the varying exogenous factors for municipalities<br />

(features of structure in combination with legal regulations), there<br />

are large differences between municipalities in spending per<br />

inhabitant.<br />

Figure 4 illustrates a survey with the spending per inhabitant per<br />

compartment (cluster) for a number of municipalities.<br />

Remarks about figure 4<br />

Figure 4 shows that:<br />

o the volume of the general means is not related to the number of<br />

inhabitants. It is possible that smaller municipalities –<br />

dependent of their features of structure – receive a larger<br />

amount per inhabitant than larger municipalities (see the<br />

municipality of Winschoten versus Haarlemmermeer);<br />

o the relative shares of the compartments within the total<br />

payment can vary significantly between municipalities;<br />

o the major differences are shown within the compartment social<br />

services (as a result of differences in social structure between<br />

municipalities);<br />

o the negative criterion that compensates for the differences in<br />

property taxes capacity leads to substantial corrections<br />

between municipalities;<br />

115


o<br />

Amsterdam (as one of the four largest cities) receives an extra<br />

fixed amount for its large city tasks. Apart from that, and as a<br />

consequence of the other criteria, the payments to Amsterdam<br />

are also rather high compared to those made to other<br />

municipalities (for a large part due to the social structure).<br />

Figure 4. The composition of the general grants in four<br />

municipalities (in euro per inhabitant)<br />

4.3. Shortcomings observed in the nineties and the search for a<br />

new method<br />

4.3.1. Shortcomings<br />

Since the 1990s the criteria (indicators) applied in the municipal<br />

fund have been selected by a different method than the one used<br />

before.<br />

This new method is directly related to the shortcomings observed<br />

in the nineties by the then ‘Raad voor de Gemeentefinanciën’<br />

(Municipal Finance Council), nowadays the ‘Raad voor de<br />

Financiële Verhoudingen’ (Financial Relations Board), regarding<br />

the allocation of the general grants of the municipal fund and the<br />

opportunities for revision.<br />

116


At that time the following was observed:<br />

o the allocation of the municipal fund was imbalanced (i.e. not in<br />

keeping with the differences to be compensated for in financial<br />

needs between municipalities based on objective<br />

circumstances);<br />

o there was an unclear and implausible relation between the<br />

criteria of allocation and the financial needs. By financial needs<br />

is understood: the - objectively determined – municipal<br />

financial needs for the realisation of services (subdivided into<br />

the above-mentioned compartments and clusters), while<br />

discounting the differences in costs between municipalities<br />

based on exogenous factors. Differences in costs based on<br />

exogenous factors are in particular connected with differences<br />

in social, physical and regional structures between<br />

municipalities.<br />

o these failings were observed shortly after a recent revision of<br />

the municipal fund;<br />

o there was no method to correct the shortcomings observed, in<br />

particular concerning the issue of the acknowledged ‘chickenegg’-problem:<br />

an imbalanced allocation has an effect on<br />

municipal spending, in which event this spending no longer<br />

serves as a proper basis for tracing differences (by means of<br />

econometric methods of research) in financial needs between<br />

municipalities as a result of features of social, physical and<br />

regional structures.<br />

4.3.2. Imbalanced growth of the municipal fund: features of social<br />

and regional structure not properly incorporated<br />

The imbalanced growth of the municipal fund occurred because<br />

features of social structure, and to a lesser extent of regional<br />

structure, were not properly incorporated in the system:<br />

o<br />

There were large dissimilarities between municipalities in the<br />

social structure of the population (low incomes; migrants). The<br />

economic stagnation in the late eighties/early nineties resulted<br />

in larger unemployment for, in particular, people with a low<br />

income or migrants. As a consequence of the varied spread of<br />

these population groups across different municipalities, the<br />

financial consequences in regard of income support, the support<br />

for employment, and all different kinds of services for social<br />

care varied similarly between municipalities. The allocation of<br />

117


o<br />

the municipal fund of that time hardly made any allowance for<br />

such differences.<br />

There were also differences in the composition and the level of<br />

regional services between municipalities (art, sports and<br />

entertainment), which were not properly incorporated in the<br />

system: the larger municipalities had to finance services for<br />

people from outside the municipalities for which they were only<br />

partly compensated in the municipal fund. Neighbouring<br />

municipalities were able to benefit from this situation as ‘free<br />

riders’.<br />

4.3.3. Origin of imbalanced growth: the use of statistic/econometric<br />

techniques for the selection and weighing of criteria (indicators)<br />

The imbalanced growth was even more remarkable in the light of<br />

the revision that the system had recently undergone (within the<br />

framework of the Allocation of Finances Act 1984). The origin of<br />

the imbalanced growth turned out to be the method for defining<br />

the weights and their criteria which had been used until then.<br />

These criteria, along with their corresponding weights, were based<br />

on statistical-economic techniques (regression). This method<br />

turned out to have serious shortcomings.<br />

1. No proper cost-oriented connection<br />

The criteria selected by way of econometric techniques did not have<br />

a proper cost-oriented connection (they did not have the proper cost<br />

drivers) with differences in spending based on exogenous factors<br />

between municipalities.<br />

2. The so-called chicken-egg problem<br />

The imbalanced allocations were actually reproduced because of<br />

regression techniques and correlations based on municipal levels of<br />

spending: the so-called chicken-egg problem.<br />

3. Only technical statistical links, as opposed to plausible links<br />

Only technical statistical links were established, as opposed to<br />

plausible links in connection with the dynamics of the system. The<br />

means towards social services, for instance, were allocated<br />

according to physical criteria (such as the number of homes). The<br />

economic changes referred to, with large differences in financial<br />

consequences between municipalities with a varying social<br />

structure as a result, were not expressed in these (physical)<br />

118


criteria. In consequence, the municipal fund did not respond<br />

adequately to the changed circumstances.<br />

4. Differences in the potential to generate own incomes<br />

The problem became even worse because of the differences in the<br />

potential among municipalities to generate their own incomes, in<br />

particular those from property taxes.<br />

4.3.4. Response to the observed shortcomings<br />

The response to the determined imbalanced growth by the then<br />

‘Raad voor de Gemeentefinanciën’ (Municipal Finance Council) was<br />

a request to a number of universities and institutions to develop an<br />

alternative research method to compensate for the shortcomings<br />

observed.<br />

In the end, the council commissioned the development of three<br />

methods for the revision of the allocation system:<br />

o further implementation of statistical/econometric techniques;<br />

o the definition of levels of services per task area and<br />

corresponding costs as a fundament of the system;<br />

o the elaboration of the so-called ‘difference analysis’, in respect<br />

of ‘wide target’ tasks areas (clusters of services) specifically<br />

focusing on the most significant cost drivers resulting from<br />

exogenous circumstances.<br />

Based on the comparison of the above three methods, the method<br />

of analysing differences, developed by Cebeon, was chosen for<br />

reassessing the municipal fund.<br />

4.4. The choice for a new method: the analysis of differences<br />

4.4.1. Why analyse differences<br />

As mentioned above, the method of analysing differences,<br />

developed by Cebeon, was chosen for reassessing the municipal<br />

fund as an answer to the imbalanced allocation of this fund,<br />

shortly after revision.<br />

The analysis of differences can be described as an iterative process<br />

attempting to explain differences in costs related to exogenous<br />

factors among municipalities with varying and similar structural<br />

features.<br />

119


The results of the analysis are:<br />

o a proper cost-oriented connection in the municipal fund<br />

between the selected indicators (in the role of cost drivers) and<br />

the differences in financial needs between municipalities;<br />

o plausible links in a dynamic sense between the selected<br />

indicators and the development of the financial needs of<br />

municipalities in respect to the effects of changing exogenous<br />

circumstances (de facto the changes in social, physical and<br />

regional structure);<br />

o a solution for the chicken-egg problem by way of a mixture of<br />

objective substantiation and political decision.<br />

4.4.2. Features of the method<br />

4.4.2.1. Basic features<br />

The method of analysing differences has the following basic<br />

features.<br />

1. Starting point: analysis of expenditure<br />

The starting point of the analysis is the ‘analysis of expenditure’:<br />

the expenditures of municipalities are corrected for differences in<br />

definitions and registration. In practice, this always turns out to be<br />

essential because tens of per cents of the differences in expenditure<br />

may be related to dissimilar definitions and registrations.<br />

2. Unravelling differences in spending according to different backgrounds:<br />

the distinction between exogenous and endogenous<br />

factors<br />

Differences in spending between municipalities are unravelled<br />

according to different backgrounds. For these backgrounds, a<br />

distinction is made between differences in expenditure (costs) in<br />

relation to municipal exogenous factors (distinguished features of<br />

social, physical and regional structure combined with the<br />

implementation of legislation) and municipal endogenous factors<br />

(policy, priorities, organisation).<br />

This process of unravelling differences in spending according to<br />

exogenous and endogenous factors is an iterative process among<br />

groups of municipalities with varying and more homogenous<br />

structural features.<br />

120


3. Rewarding differences in costs only related to exogenous factors:<br />

the solution of the chicken-egg problem<br />

The differences in costs related to exogenous factors qualify in<br />

principle for being rewarded. As ‘cost drivers’ they have to be<br />

expressed in indicators of the municipal fund. This does not apply<br />

to differences in costs related to policy and organisation.<br />

As a result, not all expenditures of municipalities follow the chosen<br />

criteria (indicators; weights) of allocation of the municipal fund,<br />

but only those expenditures related to exogenous factors. If the<br />

expenditures of a municipality are generally related to exogenous<br />

cost drivers, they are a useful base for the selection of allocation<br />

indicators. If the expenditures of a municipality are largely related<br />

to endogenous factors, they are not, or only after a correction, used<br />

for this purpose. This is the crux of the solution for the chicken-egg<br />

problem that ends in an objectified fundament of the equalised<br />

allocation of needs.<br />

4. Objective substantiation and political decisions<br />

Last, but not least important, is the relevance of political policies.<br />

The ‘Difference Analysis’ explains the background of relevant<br />

exogenous factors and indications for the relevance of more or less<br />

highly target-oriented indicators. In this manner it provides an<br />

objective substantiation of the relevant different levels of costs<br />

between municipalities, related to exogenous factors.<br />

Politics decide whether and which factors will be chosen, for<br />

example in respect to:<br />

o the choice between more global or highly target-orientated<br />

indicators;<br />

o the choice of indicators for exogenous factors that are<br />

concentrated in only a limited number of municipalities;<br />

o the choice between the use of the indicators of the Municipal<br />

Fund or the use of autonomous municipal incomes related to<br />

property value (in general of in connection with some specific<br />

services);<br />

o the introduction of new indicators (due to the decentralisation<br />

of tasks);<br />

o in the case of decentralisation of tasks: the length of the period<br />

of transition in which municipalities have to accommodate the<br />

121


122<br />

level of ‘historical’ costs to the new objective level of the general<br />

grant.<br />

In chapter 5 a number of examples of these choices will be<br />

provided.<br />

4.4.2.2. Other considerations for the design of indicators<br />

For the selection of the allocation indicators within the municipal<br />

fund, other criteria or rules are also employed in addition to a<br />

sound cost orientation following the importance of exogenous<br />

factors.<br />

A global system of allocation<br />

The analysis of differences is rather global: it is applied to ‘wide<br />

target’ connected task areas, i.e. the existing clusters in the<br />

municipal fund (see chapter 2), and a selection of a limited<br />

number of explaining variables is always made.<br />

In order to provide a clear and not too complicated system of<br />

allocation, various connected factors are generally clustered in just<br />

one criterion, with the restriction that these criteria have to meet a<br />

number of technical requirements before they can be applied as a<br />

criterion for allocation.<br />

The level of target-orientedness<br />

The criteria must have a global relation with the financial needs of<br />

municipalities (also in relation to time) and should not focus too<br />

heavily on targets. These criteria must always be balanced (see<br />

chapter 5 for a few examples).<br />

The volumes of the clusters are connected with the collective<br />

spending of municipalities<br />

The volume of the clusters is always connected with the collective<br />

spending of municipalities (as the expression of a joint, collective<br />

priority).<br />

‘Hard’ versus ‘soft’ tasks<br />

When implementing the criterion ‘cost-oriented' in relation with<br />

exogenous factors, it is important whether the expenditure is ‘hard’<br />

and required by law or whether the expenditure is largely<br />

autonomous. Examples of ‘hard’ spending are the income<br />

supplement from the social services and the means for education.


For these examples, the chosen indicators must have a close<br />

connection with differences in costs between municipalities as a<br />

result of exogenous factors (including the effects of the economic<br />

conjuncture and structure in regard with the needs of social<br />

security; or the presence in municipalities of schools for secondary<br />

and special education). For the tasks involving a high degree of<br />

autonomy in spending there is less need for a close cost orientation.<br />

An important example of the latter is the cluster ‘public parks and<br />

gardens’.<br />

Cost orientation in a dynamic sense<br />

An important condition for the allocation criteria is their accuracy<br />

in keeping up with the changes in financial municipal needs, also<br />

in relation to time. A plausible connection must therefore have<br />

been observed between the development of financial needs and the<br />

changes of the criteria due to exogenous factors.<br />

Technical quality<br />

First there are several criteria for the technical quality of the data:<br />

the data should be objectively measurable and generally available.<br />

This quality is monitored because most of the criteria originate<br />

from the central office for statistics (CBS).<br />

4.4.3. Analysis of differences as an iterative process of investigation:<br />

the example of domestic care<br />

Below is a description of the search process with an example of the<br />

cluster ‘domestic care’ for old and disabled persons). This cluster<br />

has recently been decentralised to municipalities for which the<br />

means were added to the municipality fund in 2007.<br />

1. Analysis of expenditure<br />

After identifying the relevant expenditure related to domestic care<br />

(analysis of expenditure), the backgrounds of the differences in<br />

actual spending were investigated. The analysis was made in two<br />

different ways.<br />

2. Comparison between ‘homogenous’ groups of municipalities with<br />

varying features of structure<br />

The starting point is the supposed similarity of spending among<br />

municipalities having similar structural features. At the first stage<br />

of analysis, the differences in spending were investigated between<br />

123


‘homogenous’ groups of municipalities, in this example the group of<br />

municipalities with a large extent of ageing population versus the<br />

group with a relatively younger population (examples in other<br />

clusters may be municipalities with a solid social structure or a<br />

good condition of the soil versus the group of municipalities with a<br />

weak social structure or a bad condition of the soil). By comparing<br />

the levels of spending between homogenous groups of<br />

municipalities, the influence of the distinguished varying features<br />

of structure can be quantified.<br />

3. Analysis of differences within homogenous groups of<br />

municipalities with similar features of structure<br />

Next the differences within homogenous groups of municipalities<br />

were analysed, for instance within the group of highly aged<br />

municipalities. These differences in spending may be caused by<br />

additional features of structure. By comparing different<br />

homogenous groups, it is possible to test the plausibility of a<br />

possible additional feature of structure as an explanation of the<br />

differences in spending.<br />

4. Exogenous circumstances and other explanatory factors<br />

The analysis of differences does not imply that all the differences<br />

in expenditure between municipalities are or must be related to<br />

exogenous factors (features of structure). Alternative explanations<br />

for the differences within a homogenous group of municipalities<br />

are the incidental fluctuation of spending in an individual<br />

municipality as well as regional practices of implementation<br />

(before the decentralisation to municipalities, domestic care was<br />

carried out by regional care centres). The effect of regional<br />

implementation practices can be confirmed by looking at the<br />

results of municipalities that have different homogenous groups<br />

but are located in the same region (i.e. with similar bodies of<br />

indication, care centres or care providers) compared to the average<br />

of the homogenous group.<br />

An example: the level of spending of municipality X (highly aged<br />

population, low average income) and of the neighbouring<br />

municipality Y (with a relatively young population and a high<br />

average income) is low in comparison with the average of their own<br />

homogenous groups. X and Y use the same care provider. Other<br />

neighbouring municipalities with another care provider do not<br />

124


have such a relatively low level of spending. This indicates the<br />

influence of the care provider in X and Y. This influence can be<br />

verified by obtaining information about the mode of operation of<br />

the involved care provider, which may show, for example, that<br />

poorly qualified domestic workers are assigned or that relatively<br />

few hours are provided per client.<br />

5. Iterative process<br />

The abovementioned steps of research were repeated several times<br />

until the research process resulted in an allocation formula with a<br />

restricted effect of re-allocation for homogonous groups and with<br />

effects of re-allocation for individual municipalities in relation to<br />

reasons not to be awarded (in particular differences in the way<br />

work is carried out).<br />

After the first round, a rough model of explanation was available<br />

for structural features along with global weights.<br />

By repeatedly investigating the remaining differences by means of<br />

supplementary structural features, this model could be refined<br />

both as regards the relevant structural features and as regards the<br />

weights. At this stage, the differentiation between cohorts of ages<br />

(84) was added, among other factors, and the<br />

weights were outlined. This procedure was continued until no<br />

municipality could be identified with an incomprehensible outcome<br />

(i.e. with differences in spending without a connection with<br />

structural features, executing practices or incidental fluctuation of<br />

spending).<br />

4.5. Examples of an improved connection with local needs for<br />

expenditure using the analysis of differences<br />

4.5.1. Revision and maintenance of the municipal fund by means of<br />

the analysis of differences<br />

In 1997, approximately two thirds of the municipal fund was<br />

revised, on which occasion the indicators and weights were defined<br />

not by means of techniques of regression but on the basis of the<br />

‘Difference Analysis’. In 2001, the remaining third underwent a<br />

revision in the same way.<br />

125


Large changes to the municipal fund – related to the<br />

decentralisation of new tasks and means, for instance – have also<br />

been made using the differences analysis.<br />

The method of the differences analysis was also applied for the<br />

allocation of the financial means for a number of wide targetspecific<br />

grants. This does not apply to all the specific grants.<br />

The differences analysis has resulted in a better connection<br />

between the selected indicators and the differences in local needs,<br />

also if exogenous factors were only relevant for a limited number of<br />

municipalities. The plausible connections are also made in a<br />

dynamic sense. A clear cost orientation is incorporated into the<br />

allocation system – also in a dynamic sense – for each<br />

compartment and for each cluster. Incorporated differences in costs<br />

are related to exogenous factors, based on features of social,<br />

physical and regional structure.<br />

As a result, the indicators of the Municipal Fund form an<br />

‘automatically working system’: the system works with relatively<br />

limited efforts of maintenance (see chapter 6) and without<br />

imbalances in the long term.<br />

4.5.2. Examples<br />

Next, a number of examples are presented of the more explicit and<br />

plausible connection between the system of allocation and local<br />

financial needs by means of the differences analysis as well as of<br />

political policies.<br />

Differences in costs due to differences in conditions of the soil<br />

The system of indicators of the Municipal Fund consists of<br />

indicators for the allocation of differences in costs related to<br />

differences in the condition of the soil (subsiding peaty soil). Such<br />

criteria could not be determined by way of regression techniques,<br />

because the problem is concentrated in a limited number of<br />

municipalities, and the expenditures are hidden in a variety of<br />

clusters.<br />

Politicians became involved in the decision whether the Municipal<br />

Fund had to compensate for higher costs in cases where houses<br />

and roads had been constructed on subsiding peaty soil. The<br />

126


answer was positive due to the influence of national environmental<br />

planning on this subject.<br />

Because of the municipalities’ own responsibility for building on<br />

this type of soil, the Municipal Fund does not give a full<br />

compensation for all kinds of costs.<br />

Introducing a number of highly target-oriented indictors<br />

The indicators for ‘students in special and secondary schools’ are<br />

introduced in the Municipal Fund. Education is incorporated<br />

because of the fact that this kind of schools (and costs) are<br />

unevenly spread across municipalities and are not expressed in<br />

other – more global - criteria. Besides, the decision to establish a<br />

school is not exclusively made by municipalities, but also by other<br />

bodies. As a result, the spread of specific and secondary schools is<br />

mainly an exogenous factor for municipalities.<br />

The same argument holds for the use of the highly target-oriented<br />

indicator ‘persons drawing benefits’. The number of these benefits<br />

is strongly related to changes in economic conjuncture and<br />

structure in combination with differences in social structure<br />

between municipalities. These factors are for the major part<br />

exogenous for municipalities.<br />

Introducing a number of more global working indicators<br />

Related to other clusters (such as art, sports, entertainment and<br />

public parks and gardens), more global working indicators (in the<br />

form of the number of inhabitants of houses) are introduced<br />

instead of more target-oriented indicators such as the areas of<br />

roads and parks or the presence of a theatre.<br />

The differences analysis on the cluster ‘Public parks and gardens’<br />

reveals that the volume and composition of spending on public<br />

parks and gardens can vary extensively between municipalities.<br />

This is partly related to a rural structure (with a lot of extensive<br />

forms of green) or a more urban one (with less but intensively used<br />

parks and gardens), and partly it is obvious that the actual<br />

implementation is strongly related to the individual policy<br />

preferences of municipalities. For this reason, the choice was made<br />

to allow a global manner of allocation of the financial means for<br />

127


public parks and gardens through the indicators inhabitants and<br />

housings.<br />

Choices related to the costs of the fire brigade<br />

In several communities (in particular the smaller ones), the Dutch<br />

fire brigades are manned by volunteers. Other communities (in<br />

particular the larger ones) have their own professional fire brigade.<br />

The indicators of allocation of the Municipal Fund reflect these<br />

differences globally.<br />

In recent years, the costs of fire brigade and public security have<br />

been rising for different reasons. In respect to the allocation of<br />

extra financial means, a choice is made between the use of the<br />

indicators of the Municipal Fund and the use of the autonomous<br />

municipal incomes related to property value.<br />

No compensation in the Municipal Fund for exemptions<br />

To a certain degree municipalities are authorised to grant<br />

exemption from local taxes and charges.<br />

Because of the municipal responsibility in this matter, politicians<br />

decided not to compensate these exemptions by way of the<br />

Municipal Fund.<br />

The introduction of several indictors related to different kinds of<br />

geographical and building structure<br />

The differences analysis explains and makes explicit the<br />

background for differences in costs related to different kinds of<br />

geographical and building structures of municipalities. As a result,<br />

several indicators have been introduced, such as multiple centrism,<br />

the historical centres and the historicity and volume of internal<br />

and external waters.<br />

Indicators for new municipalities<br />

The reference dates of indicators are in general one or two years<br />

behind reality. Most of the municipalities have not had (financial)<br />

difficulties concerning this matter, except for new municipalities<br />

with a fast growing population. These (financial) difficulties for<br />

new municipalities are recognised by way of the differences<br />

analysis, which led up to a form of compensation in the Municipal<br />

Fund.<br />

128


Correcting endogenous factors related to differences in costs of<br />

domestic care<br />

In chapter 3, the process of ‘Differences Analysis’ was described for<br />

the costs of domestic care (for old or disabled persons). This<br />

analysis made it clear that cost differences for tens of per cents are<br />

related to endogenous factors such as systems of indications or the<br />

chosen levels of services.<br />

Politicians decided to compensate only for real exogenous factors<br />

(like the number of old, disabled persons) and not for the<br />

abovementioned endogenous factors.<br />

Difficulties with systems of allocation for specific grants, design<br />

with regression techniques<br />

Not all wide target-specific grants are designed using the<br />

differences analysis. An important example is the system of<br />

allocation for the specific grant for ‘Work and income’. This specific<br />

grant has been developed using regression techniques. With<br />

respect to this specific grant we see large difficulties in introducing<br />

a cost-oriented, plausible and ‘automatically working’ system of<br />

indicators. The effects of these difficulties are annual adaptations<br />

with large effects of reallocation and a continuing discussion (also<br />

with the government) with respect to the working of the system.<br />

4.6. The institutional environment of the Municipal Fund<br />

Thanks to the differences analysis, the equalised allocation of local<br />

expenditure needs by means of the Municipal Fund in the<br />

Netherlands is a well-balanced mixture of an objective<br />

substantiation of local financial needs and political decisions.<br />

This mixture is realised and protected by the following<br />

institutional environment.<br />

The availability of an objectified fundament<br />

o research (Differences Analysis);<br />

o qualified data in a technical sense (CBS).<br />

Periodic maintenance and monitoring<br />

o periodic maintenance through annual scanning of the<br />

development of spending versus the allocation system, if<br />

129


o<br />

needed supplemented by more detailed research on the<br />

backgrounds of differences between the expenditures of<br />

municipalities and the working of the allocation system;<br />

monitoring the rules and quality by the administrators of the<br />

municipal fund (the Ministries of Interior Affairs and Finances)<br />

and the “Raad voor de financiele verhoudingen” (Financial<br />

Relations Board).<br />

A well-balanced decision-making procedure<br />

A well-balanced decision-making procedure for adaptations of the<br />

system:<br />

o supervision of the (results of) research by an advisory council.<br />

This council includes, along with the administrators of the<br />

municipal fund, the Association of Dutch Municipalities (VNG).<br />

The input of the bodies concerned is fundamental to any<br />

adaptation (individual municipalities with different<br />

characteristics of social, physical and regional structure;<br />

Association of Dutch Municipalities);<br />

o a proposition for a change from the government, i.e. the<br />

Ministry of Interior Affairs;<br />

o the final decision made by the Lower Chamber of Parliament.<br />

In general the Lower Chamber only deals with the outlines of<br />

the allocation and in particular monitors the effects of<br />

reallocation and specific circumstances of certain types of<br />

municipalities (also as an effect of municipal lobbying). In the<br />

past few years, for instance, the Lower Chamber has asked for<br />

new research regarding the problems of a bad condition of the<br />

soil, of municipalities with urban growth, of rural, multiple<br />

centred municipalities and of the major cities.<br />

130


Chapter 5<br />

Assessing regional and local government<br />

expenditure needs in Italy. Small<br />

achievements and large prospective issues.<br />

Giorgio Brosioand Stefano Piperno<br />

5.1. Introduction<br />

Subnational finance in Italy has been dominated since the 1990s<br />

by the reintroduction of tax autonomy. Own revenues have to a<br />

substantial extent replaced intergovernmental grants. These<br />

grants are basically allocated according to the historical criterion,<br />

rewarding low income and wealth conditions and small size of local<br />

units. In other words, the Italian grant system is geared to protect<br />

poor and small (in terms of their population) subnational<br />

governments.<br />

As in every historically-based transfer system, inefficiency and<br />

inequity have accumulated over the years. The system has been<br />

subject to frequent changes, but not to overall reform. Attempts to<br />

introduce objective need factors have failed, and the allocation of<br />

grants has been increasingly geared to equalise revenue in a period<br />

where disparities of per capita own revenues hugely increased as a<br />

consequence of the reinvention of local tax autonomy.<br />

Surely, lack of reform is also due to weak central government<br />

guidance, huge fragmentation, large disparities between the<br />

various areas and institutional and political infighting.<br />

A new legal framework for the intergovernmental transfer system<br />

is provided by the new constitution of 2001 that introduces the<br />

concept of minimum standards (“essentials levels” in the Italian<br />

constitutional terminology) of service provision for some services<br />

that are considered as fundamental to the wellbeing of the<br />

131


population. These services are identified only for the regional level<br />

of government and include health and social protection. For local<br />

governments, the specification of these fundamental services is<br />

delegated to ordinary law. While minimum standards apply to the<br />

fundamental services, the constitution seems to mandate that<br />

equalisation for all the other functions has to be based on fiscal<br />

capacity and without consideration of expenditure needs for these<br />

functions.<br />

The paper is divided into two parts. The first part illustrates the<br />

present system and its evolution over the two last decades. The<br />

attention is focused on the attempts that have been made to<br />

introduce components based on needs. The second part presents<br />

the perspectives and the issues underlying the new constitutional<br />

discipline.<br />

5.2. The present system of intergovernmental grants in Italy<br />

5.2.1. Background<br />

Italian subnational government has been substantially reformed in<br />

the last twenty years. The steady devolution of taxation powers to<br />

local governments has gone hand-in-hand with the reallocation of<br />

spending responsibilities.<br />

In l990, sub-national governments accounted for only<br />

approximately 8 per cent of the total national general government<br />

revenue, while their share of total national government<br />

expenditure was 38 per cent. In 2006, the share of subnational<br />

government climbed to 23 per cent of the general government<br />

revenue (see table A2 in the Appendix), while its share of total<br />

national expenditure is 46 per cent, net of social security and<br />

interests on debt. To date, municipal and provincial governments<br />

make up about 20 per cent of total national expenditure, while the<br />

regions, together with the health authorities (which are offshoots<br />

of the regions), account for the 26 per cent (cf. table A3 in the<br />

Appendix).<br />

Central transfers constitute more than fifty per cent of regional<br />

revenues. They represent one third of provincial revenues and one<br />

fourth of municipal revenues (see table 1).<br />

132


Table 1. Structure of revenue of subnational governments, 1990-<br />

2005<br />

Provinces Communes Regions<br />

Revenues 1990 2005 1990 2005 1990 2005<br />

Own taxes 8.5 59.2 19.4 52.5 1.6 39.9<br />

Fees and user charges 0.5 0.9 9.8 11.9 0.1 0.1<br />

Grants 85.5 33.3 65.1 26.0 97.7 59.3<br />

Non tax revenues (*) 5.5 6.6 5.7 9.6 0.6 0.7<br />

Total 100 100 100 100 100 100<br />

Source: Ministry of Economy, Country General Economic Report, various years.<br />

Accrual revenues.<br />

(*) Net of borrowing.<br />

From 1990 to 2006, transfers from the central to the local<br />

governments decreased from 51 per cent to 22 per cent of local<br />

governments’ expenditure. At the same time, the importance of<br />

transfers from regions to local governments in terms of their<br />

expenditure increased from 11% to 17 per cent. (see table 2). This<br />

means that the vertical fiscal imbalance has been considerably<br />

narrowed.<br />

Table 2. Share of transfers from Central and Regional<br />

governments against total expenditures by Provinces and<br />

Municipalities 1990-2006<br />

Transfers from: 1990 1995 2006<br />

State 50.91 34.47 22.43<br />

Regions 10.69 10.22 16.69<br />

Source: Ministry of Economy, Country General Economic Report, various years.<br />

5.2.2. Main features of intergovernmental grants<br />

There have traditionally been two distinct general transfer<br />

systems in Italy, one for the regions, and one for the local<br />

governments (municipalities and provinces).<br />

Ordinary regions 58<br />

Initially (1970), two general funds, the Common Fund and the<br />

Regional Development Programs Fund, were introduced to finance<br />

regional expenditures. The allocation of the first fund was<br />

58<br />

Special Statute Regions (there are 5 of them) have a completely different<br />

system of financing that is mostly based on shared taxes. They will not be<br />

considered here.<br />

133


determined according to varying percentages of a number of excise<br />

taxes, while the allocation of the second was discretionary. These<br />

funds gradually lost their unconditional character and were<br />

replaced by sectoral funds, concentrated in the health and<br />

transportation sectors. The experience of the two main sector funds<br />

showed that equilibrium could hardly be found between central<br />

governance of the system (through guidelines and uniform<br />

standards, criteria for the interregional distribution of funds and<br />

checks on their use) and regional autonomy in planning and<br />

management of health structures and local public transportation.<br />

In the late 1990s, a new system for financing regions was devised.<br />

A regional tax on business activities (IRAP) and a regional surtax<br />

on personal income were introduced. The former is a direct-type<br />

value-added tax to be levied on all business activities as of 1998.<br />

The standard tax rate was set at the uniform level of 4.25 per cent,<br />

but regions are entitled to raise or lower their tax rate by a<br />

maximum of 1 per cent and to apply varying rates to different<br />

sectors and categories of taxpayers The new tax has a broad,<br />

potentially very productive tax base: value added, net of<br />

depreciation, but including interest payments. It has become the<br />

major regional tax.<br />

Secondly, the main special purpose funds were abolished and<br />

replaced by shared taxes - the personal income surtax at a<br />

standard flat rate, the petrol tax and VAT - and equalisation<br />

grants. Shared taxes – when added to their own taxes - allowed the<br />

wealthiest regions of the North and the Centre to cover all their<br />

spending responsibilities. While the petrol tax and the personal<br />

income surtax were shared on strict derivation, the VAT was<br />

allocated according to equalising criteria.<br />

After a transition period, the VAT had to be allocated among<br />

regions according to fiscal capacity and expenditure needs. 59 The<br />

intended grant system distinguished between two main sectors of<br />

expenditure reflecting: a) health needs and, b) non-health needs.<br />

Health expenditures represented 72 per cent of the regions’ total<br />

expenditures.<br />

59<br />

For details see the Appendix.<br />

134


a) Health needs. These needs were introduced separately because<br />

of the central government mandate to introduce essential levels of<br />

health care.<br />

Assessment of health needs distinguished between six categories of<br />

essential services (LEA). The age structure of the population was<br />

used as the crucial indicator of need. (cf. table 4). The basic<br />

formula is summarised as follows:<br />

Ria=Pij.U * aj.S * a. ( 3 √Bi only for hospitals)<br />

With:<br />

i= 1-15 (Ordinary Statute Regions)<br />

a= 1-6 (categories of Essential levels of Health Care – LEAs)<br />

j = age groups, and<br />

where Ria indicates the share of the total allocation fund assigned<br />

to region i for the LEA a, U * aj indices of health consumption by age<br />

for the different LEA (assumed to be equal in all regions), S * a the<br />

average per capita expenditure for the LEA a at the national level,<br />

and Bi is the standardized mortality rate in region i.<br />

The use of the age structure of the population is standard in the<br />

literature on health care and in actual government practice. This<br />

method aims at transforming the effective population of every<br />

region into a “virtual” population, taking into account the<br />

differences in demand for the health care services within different<br />

demographic groups.<br />

For instance, the weights used in Italy with regard to the hospital<br />

LEAs amount to the following factors:<br />

2.539 for individuals less than one year old;<br />

0.254 for individuals between the 5-14 years range;<br />

3.025 for individuals older than 75 years.<br />

In this way, the equivalent expenditure per capita can be defined<br />

as the total health care expenditure divided by the weighted<br />

population, rather than by the effective population. 60<br />

60<br />

In the recent years, however, the formula for assessing the relative health<br />

expenditure needs among the regions has been simplified. The latest version<br />

135


Table 3. Essential levels of health care (LEAs). Relative weights<br />

and need indicators for 2006<br />

Essential levels of Health Care (LEAs)<br />

Amount of<br />

the Fund (%)<br />

Indicators of health<br />

needs<br />

Public health promotion and<br />

prevention programmes 5 Population<br />

Outpatient care 51<br />

° Pharmaceuticals 13<br />

population weighted by<br />

age group<br />

° Ambulatory services 13<br />

population weighted by<br />

age group<br />

° Other outpatient health services 18.1 Population<br />

° General practitioners 6.9 Population<br />

Inpatient care (hospitals) 44<br />

population weighted by<br />

age group<br />

Source, Ministry of Health 2006.<br />

b) The details of non-health expenditure needs calculation are<br />

given in the Appendix. <strong>Needs</strong> were estimated with the help of a<br />

regression model, which assessed the extra costs due to the small<br />

dimension of some regions. This effort to introduce a system based<br />

on needs has been frustrated by the opposition of some, less<br />

developed, Southern regions, when they realised (only at the<br />

moment the system was being implemented) that they were going<br />

to lose from it. Thus, the new allocations have been frozen and the<br />

system continues to be based, for non-health expenditures, on the<br />

previously existing patterns based on bargaining.<br />

Health care expenditure is dominant, and the system of essential<br />

levels ensures uniformity of per capita resources to all regions,<br />

while sector transfers (Giarda, 2005), mainly for capital purposes<br />

and tied to EU regional structural funds, ensure that, in total, the<br />

poorest regions have per capita expenditure levels higher than<br />

those of the richest regions (see figure 1).<br />

(2006) is based for 66 per cent on weighted capitation shares and for the rest on<br />

simple capitation shares.<br />

136


Figure 1. Per capita GDP and per capita expenditures of Ordinary<br />

statute regions. 2005<br />

2500<br />

2000<br />

health expenditures<br />

Pper capita expenditures<br />

1500<br />

1000<br />

non health expenditures<br />

500<br />

0<br />

14000 16000 18000 20000 22000 24000 26000 28000 30000 32000 34000<br />

per capita GDP<br />

Source: Audit Commission,2007<br />

In fact, the allocation of grants to regions seems to be dominated<br />

more by motives of convergence of regional development levels<br />

than by equalisation of service delivery levels. The latter would<br />

require, at most, equal per capita levels of expenditure and not a<br />

negative correlation between these levels and the GDP (table A4<br />

and A5 in the appendix).<br />

Municipal and Provincial <strong>Government</strong>s<br />

The system for allocation of general grants to local government is<br />

and has been, as in the case of regions, rather obfuscated. Its<br />

complexity challenges a clear and concise description. It has been<br />

subjected over the years to frequent changes that have not altered<br />

its fundamental characteristics. The system of allocation favours<br />

the small-sized local government units and the large ones and has<br />

no clear equalising impact.<br />

• Until 1993 a large quota of transfers was targeted at<br />

equalisation, though the criteria used varied over time. The<br />

137


main objectives were: evening out differences in per capita<br />

expenditure of municipalities of similar sizes;<br />

• larger per-capita transfers were given to small and large<br />

municipalities. The assumption was that unit costs of local<br />

governments were U shaped;<br />

• municipalities that were poor and/or situated in poor areas<br />

received higher per-capita transfers.<br />

This policy was considered by the Ministry of the Interior, which<br />

was in charge of it, as a success, since it reduced to a considerable<br />

extent the variance of per capita spending of municipalities of<br />

similar size. In reality, the policy was largely questionable in terms<br />

of both efficiency and equity. This is because municipalities with<br />

the same population size have actually different production costs<br />

and needs due to the physical setting, climate, density of<br />

population, commuting, etc.<br />

Table 4. Method for assessing expenditure needs in 1994-95<br />

Basic services<br />

Need/cost factor measurement<br />

unit<br />

Population<br />

groups<br />

General administration and<br />

justice (*)<br />

Population 0-499<br />

inhabitants<br />

<strong>Local</strong> police Population 500-999<br />

Garbage collection and<br />

disposal<br />

weighted urban area plus non<br />

urban area in km 2 1000-1999<br />

Primary education<br />

share of population between 6<br />

and 10 years 2000-2999<br />

Secondary education<br />

share of population between 11<br />

and 13 years 3000-4999<br />

Cemeteries<br />

population weighted with life<br />

expectancy 5000-9999<br />

Streets and public lighting<br />

weighted urban area plus non<br />

urban area in km 2 10000-19999<br />

Water urban area in km 2 20000-59999<br />

Sewage urban area in km 2 60000-99999<br />

<strong>Expenditure</strong> needs of each municipality is assessed multiplying<br />

average expenditure per measurement unit in each population<br />

class by its actual measurement unit for each service. This value<br />

was adjusted with a deprivation indicator and with a weighting<br />

100000-<br />

249999<br />

250000-<br />

499999<br />

factor tied to the presence of militaries.<br />

(*) includes five individual services. Beyond<br />

500000<br />

138


Act n.142/90 reforming local government introduced a new system<br />

of grants for municipalities based largely on expenditure needs and<br />

revenue capacity and was implemented in 1992. The new system<br />

was necessitated, among other things, by the introduction of the<br />

local property tax. This tax is levied by municipalities, which have<br />

the power to fix the tax rates between relatively wide centrally<br />

determined brackets. At the end of the nineties, local governments<br />

were also assigned a local surcharge on personal income tax.<br />

The new system had three distinct funds for recurrent expenditure:<br />

a) the Recurrent Fund<br />

b) the Equalization Fund, and<br />

c) the Conditional Fund.<br />

These funds were supplemented by two additional funds for capital<br />

expenditure, based mostly on sector legislation, such as that for<br />

school building and urban infrastructure.<br />

The Recurrent Fund was aimed at financing a minimum and<br />

uniform level of basic services.<br />

For allocation purposes, it used a number of indicators specific to<br />

the distinct municipalities - such as the population and its age<br />

structure, the size of urban and extra-urban areas, population<br />

density and access of population to services.<br />

Some other indicators were referred to the socio-economic<br />

conditions of the province where municipalities are located. Table 6<br />

summarizes this method. As reported, the main determinant was<br />

population. The basic assumption behind its use was again that<br />

unit costs of local governments are U-shaped and that any transfer<br />

system had to replicate this distribution. To this aim,<br />

municipalities were and still are divided into twelve different<br />

classes of population. For each class, the grant per capita is<br />

calculated with reference to a uniform level of service provision.<br />

Then, this amount is adjusted by using specific municipal and<br />

provincial indicators.<br />

139


Formally the measurement unit cost for each basic service is:<br />

n n<br />

PMO 1 j= ∑Si/∑UDji<br />

i=1 i=1<br />

With:<br />

Si= current expenditure of Commune in population class 1 for the basic service<br />

j<br />

UDji= measurement unit cost for the service j of Commune i<br />

<strong>Expenditure</strong> needs for the service j in the individual Commune i of class 1 is<br />

given by:<br />

FABBij= PMO 1 j*UDji<br />

While the sum of the expenditure needs for all basic services in population class<br />

1 is given by:<br />

14 14<br />

∑ FABBij=∑ PMO 1 j*UDji<br />

j=1 j=1<br />

The total of expenditure needs of all Communes will be given by:<br />

n 14 12<br />

∑ ∑ ∑ FABB z ij<br />

I=1 j=1 z=1<br />

The total amount of the Recurrent Fund was to be fixed in nominal<br />

terms, while its increases, in line with the inflation, ought to have<br />

been added to the Equalisation Fund. This Fund had to be<br />

distributed in such a way as to gradually equalise per capita tax<br />

revenues of each local unit by bringing them close to the average<br />

calculated for all units included in the same population bracket.<br />

One of the main problems in implementing this mechanism was<br />

that, since information of effective tax bases was basically<br />

inadequate, actual collections were used as a proxy of revenue<br />

capacity, giving an incentive to lower tax effort.<br />

Table 5. Components of the grants system for the municipalities<br />

1994 % 2006 %<br />

Recurrent Fund 13577 76 9021 81<br />

Equalisation Fund 742 4 773 7<br />

Conditional fund 3625 20 1300 12<br />

Total 17944 100 11094 100<br />

Source: Home Affairs Ministry. 1994, billions of lire; 2006 millions of euro.<br />

140


The working of this system (Ministero dell’Interno, 1995) has been<br />

frozen after 1995 due to strong disagreement among<br />

municipalities, and individual allocations have been adjusted<br />

annually in a totally non-transparent way, which defies efforts of<br />

researchers, but seems to please most local government units,<br />

since no request for reform has been loudly advanced.<br />

Table 6 shows the U-shaped curve of per capita municipal<br />

expenditure, to which the policy of size-differentiated transfers has<br />

contributed fundamentally.<br />

Table 6. Per capita recurrent expenditure of Italian municipalities<br />

by population size. 2001 and 2005<br />

Per capita<br />

Per capita<br />

Population classes<br />

Index.<br />

Index.<br />

recurrent<br />

recurrent<br />

Smallest<br />

Smallest<br />

expenditure<br />

expenditure<br />

Class = 100<br />

Class = 100<br />

(000 of lires)<br />

(000 of lires)<br />

2001<br />

2005<br />

Less than 500 1,055.6 100 1,214.8 100<br />

500-999 993.0 86 1,158.9 96<br />

1000-1999 735.2 64 878.8 73<br />

2000-2999 676.7 59 741.4 61<br />

3000-4999 737.4 64 694.7 57<br />

5000-9999 647.7 56 642.7 53<br />

10.000-19,999 680.9 59 678.9 56<br />

20,000-59,999 737.5 64 723.5 60<br />

60,000-99,999 759.1 66 831.8 69<br />

100,000-249,999 891.8 77 844.6 70<br />

250,000-499,999 1,104.4 96 914.2 76<br />

More than 500,000 1,232.9 107 1,256.2 104<br />

Source: Ministry of Economy, Country General Economic Report, various years<br />

5.2.2 Main weaknesses of the system<br />

A) Regional governments<br />

141


Health needs:<br />

• The assessment of needs tied to LEAs is mainly based on<br />

population. As a consequence, per capita expenditure in<br />

health services is virtually equalised across regions. Despite<br />

equal levels of expenditure, huge disparities prevail in<br />

effective outcomes, as proven by the uninterrupted<br />

interregional migration of patients from the Southern<br />

regions to the Central and Northern regions in “search of<br />

quality”.<br />

• To properly assess health expenditure needs, indicators of<br />

health risk should be taken into account in addition to<br />

population indicators. This is because the age structure of<br />

the population only partially represents the health care<br />

requirements of individual regions. Other variables<br />

(environmental, cultural and socio-economic factors) are<br />

relevant to the determination of the effective regional<br />

healthcare requirements.<br />

• The central government practice of bailing out the Health<br />

authorities, whose budget deficit originates from spending<br />

surpluses, has muted the essence of the whole mechanism.<br />

In practice, health expenditure needs are re-determined<br />

annually (ex post) through painstaking negotiations<br />

between the regions and the central government.<br />

Non health needs:<br />

• Taking only size into account legitimises existing scale<br />

diseconomies. This implies and maintains a redistribution<br />

of resources from the largest to the smallest regions, which<br />

are not necessarily the poorest.<br />

B) Provinces and municipalities<br />

• Size in terms of population has always been the dominant<br />

factor in assessing the expenditure needs with regards to<br />

“basic services” that - according to official estimates<br />

(Ministero dell’Interno, 2002) - represent about 57 per cent<br />

142


of communes’ expenditures and 67 per cent of provincial<br />

expenditures.<br />

• However, the “cost of fragmentation” is relatively small. In<br />

Italy, today more than half (57 per cent) of the<br />

municipalities have less than 5000 inhabitants, but their<br />

share of total municipal expenditures is only 16 per cent.<br />

• The issue of specific needs of largest municipalities<br />

(metropolitan areas) has never been addressed.<br />

5.3. Towards a new system<br />

5.3.1. The new discipline of grants<br />

With the reform of 2001, transfers to sub-national governments<br />

have been subjected to constitutional discipline. The constitution<br />

introduces three different types of transfer. The first one is<br />

implicitly introduced via the constitutional provision that<br />

mandates the central government to determine minimum<br />

standards (“essential levels”) of service provision applying to<br />

fundamental responsibilities, such as health and social assistance,<br />

and to ensure their provision via appropriate financing. This<br />

requires a system of transfers based on minimum expenditure<br />

needs. Recent national legislation has extended to a still<br />

undefined set of municipal and provincial functions the domain of<br />

fundamental responsibilities, to which minimum standards will<br />

apply 61 .<br />

The second set of equalisation transfers refers to the remaining<br />

regional and local expenditure functions. Here, regional and local<br />

government units with a fiscal capacity below the national average<br />

will be entitled to equalising grants based on the fiscal capacity<br />

only.<br />

The third set of grants includes grants from EU programmes and<br />

the corresponding Italian co-financing transfers. It includes also<br />

specific grants from the central government targeted at filling<br />

regional disparities in growth. However, the aim of the new<br />

Constitution was to limit as much as possible the use of this kind<br />

61<br />

We may assume that they include the former “basic services”.<br />

143


of transfer. More precisely, the Constitution states that<br />

equalisation grants must all be unconditional, while specific grants<br />

are allowed only when allocated to subnational governments with a<br />

low level of economic development and in addition to their ordinary<br />

resources. This constitutionally mandated reduction of sectoral<br />

transfers can be ascribed to the increased awareness of the risks<br />

associated with widespread use of this instrument, such as<br />

excessive bargaining between individual layers of government with<br />

huge delays in the determination of the amount to be transferred;<br />

inefficient budgeting processes; regional overspending, growth of<br />

subnational deficits 62 . Therefore, conditional grants will have to be<br />

significantly reduced. For example, existing grants for<br />

kindergartens have been discontinued 63 due to an in-depth<br />

interpretation of the new constitutional text by the Constitutional<br />

Court.<br />

In the future, the Italian transfer system will have to be almost<br />

completely based on the former two types of formula-driven<br />

equalisation grants.<br />

5.3.2. The stakeholders<br />

The municipalities<br />

There are more than 8,000 municipalities in Italy. Their sizes, in<br />

terms of population, vary from less than 50 to more than 3 million<br />

inhabitants. The size distribution of Italian municipalities still<br />

reflects the one that prevailed in the pre-unitary States (that is,<br />

the situation preceding the creation of the country). While in the<br />

Northern regions a huge fragmentation prevails, the municipalities<br />

located in the Southern regions are much larger. A typical Sicilian<br />

rural municipality has between 30,000 and 50,000 inhabitants.<br />

This number corresponds more or less to the size of the small<br />

provincial capital cities in the North. Obviously, there is a huge<br />

diversity between the two samples of municipalities in the urban<br />

functions they perform, despite the similarity in their population.<br />

The variance of income and wealth conditions is also considerable.<br />

62<br />

In the constitutional debate we have not found a clear reference to the<br />

prescriptions of the European Charter of <strong>Local</strong> <strong>Government</strong> aimed at reducing the<br />

use of grants earmarked for specific projects.<br />

63<br />

While the pending assignments have been preserved<br />

144


The average per capita tax base of the municipalities of Calabria -<br />

the poorest region - is barely one third of that of the richest one,<br />

namely Lombardy. Also climatic conditions are extremely different,<br />

due to the diversity of latitude and altitude and of their<br />

combination.<br />

The overwhelming majority of municipalities (meaning their<br />

political appointees and their administration officials) is at best<br />

lukewarm towards the reform of the grant system. In fact, this<br />

system gives them certainty of revenue and looks more attractive<br />

than the prospect of a reform where, considering the present<br />

climate of fiscal restructuring, losses could outweigh gains. In<br />

other words, officials and politicians show a risk-averse behaviour.<br />

Comparisons with what other municipalities get do not seem to<br />

play a great role. The system has been in place for a considerable<br />

time and has acquired a kind of legitimacy.<br />

Over the years a common front against the central government has<br />

arisen that prevails against idiosyncratic differences. In other<br />

words, Italian municipalities present a rather common platform<br />

towards the issue of their grant system and its reform - a main<br />

component of this platform being a preference towards steady<br />

annual increases of the present allocations over an overhaul of the<br />

system. Growing local tax autonomy makes most local units<br />

lukewarm towards a need-based reform. The rich units benefit<br />

from an expansion of local tax autonomy, while the poorest units<br />

are mostly interested in revenue equalisation.<br />

Their association<br />

Municipalities are represented by the National Association of<br />

Municipalities (ANCI), which is subdivided into regional branches.<br />

The chairman of ANCI is traditionally elected among the<br />

representatives of small and medium-sized municipalities, mainly<br />

because of their large number. As a consequence, the association<br />

has a clear small-medium sized municipality orientation. Large<br />

municipalities mainly have a veto power inside ANCI and have<br />

direct access to the central government. Party divides do not seem<br />

to play a fundamental role within the association, both because of<br />

the bipartisan tradition of Italian politics, where government and<br />

opposition parties often cooperated on financial decisions, and<br />

because of the increasingly higher variance of electoral fortunes of<br />

145


political parties at the municipal level.<br />

The reform of the grant system does not rank high among ANCI<br />

priorities. Considering the huge disparities between individual<br />

municipalities, assembling a sufficiently large and united front for<br />

a reform would be a quite cumbersome and risky task. ANCI<br />

prefers by large to support individual requests for marginal<br />

changes to the present system and to negotiate every year with the<br />

central government a substantial increase on the previous total<br />

allocation. Most of ANCI's activity concerning municipal finances<br />

takes place on a day-to-day basis and is concentrated on asking<br />

compensation for every central decision that impacts negatively on<br />

the finances of municipalities.<br />

The central government<br />

There are two main actors playing at the central level, namely the<br />

Ministry of the Interior (MoI) and the Ministry of Finance (MoF).<br />

MoI is the traditional tutor of Italian municipalities. In the former<br />

centralised system, its main functions consisted, in addition to<br />

finance, in monitoring and control.<br />

MoI has elaborated the system for the allocation of general grants,<br />

and the Ministry is still responsible for its implementation,<br />

although the MoF makes the actual disbursements. Monitoring<br />

and control have been replaced by sponsoring. In other words, the<br />

MoI represents the interests of municipalities at the central,<br />

government level (including Parliament). Over the years, a close<br />

working relationship has been forged between MoI and ANCI. The<br />

latter prepares the requests, which are then brought by the MoI to<br />

the attention of the competent Ministry, or are translated into<br />

parliamentary bills, when the need arises.<br />

As in most countries, the MoF is keen on replacing the MoI in<br />

every aspect of the intergovernmental transfer policy. As<br />

intergovernmental grants are a considerable item in the central<br />

budget, the MoF would like to play an active role, in the hope that<br />

a more efficient system of allocation would bring savings in their<br />

total amount. Moreover, the MoF rightly believes that controlling<br />

the grant lever could ease the present difficulties of implementing<br />

the domestic Stability Pact.<br />

146


Regional governments<br />

These governments do not by far present a common front. The<br />

main dividing line is drawn by income and wealth disparities.<br />

Northern and Central regions, being much wealthier than thos in<br />

the South, are rather cold towards intense redistribution of<br />

resources through the grant system. Different political orientations<br />

play a smaller role, while there is substantial animosity between<br />

ordinary and special statute regions, these latter ones being hugely<br />

privileged in the distribution of resources.<br />

Regional governments, especially those situated in North and<br />

Central Italy, have increasingly asked to be made responsible of<br />

allocating grants to municipalities (given, of course, the provision<br />

of the corresponding finance by the central government). However,<br />

such pleas have made very little inroad in the minds of municipal<br />

officials. As is usually observed across the world, municipalities<br />

prefer to interact with a more distant payer, that is, with the<br />

central government, rather than with the regions - partly because<br />

the former is considered financially more viable and politically fair.<br />

The recent government bill implementing the constitution of 2001<br />

has partially accepted the regional requests by making regions<br />

responsible for allocating general grants to the small<br />

municipalities.<br />

Intergovernmental consultative bodies<br />

As mentioned in the introduction, Italy has developed two<br />

intergovernmental coordinating bodies. They are the Conferenza<br />

Stato-Regioni and the Conferenza Stato-Regioni Città 64 .<br />

All central government decisions and laws having an impact on<br />

subnational governments have to be presented for evaluation to<br />

these bodies. Both conferences have also to agree at unanimity on<br />

the annual allocation of grants to each level of government. Due to<br />

the incremental nature of the system of allocation, both<br />

conferences usually easily reach an agreement. The Conferenza<br />

64<br />

The former consists of representatives of the central government and of the<br />

Governors of all regions (who can appoint substitutes with reference to the issues<br />

dealt in meeting). The second conference is constituted by representatives of the<br />

central government, by the chairmen of the municipal and provincial government<br />

and Mountainous Communities associations, by 15 mayors (of which 5 are from<br />

metropolitan cities) and by 6 presidents of provincial governments.<br />

147


Stato-Regioni had also, after long lasting debates about the<br />

intensity of equalisation and the timing of the reform, reached an<br />

agreement on the reform of the transfer system for regions (Decree<br />

N. 56, 2000). However, as already mentioned, when the new<br />

system was supposed to be started in 2002, some Southern regions<br />

opposed its introduction, after realising that they were not (in<br />

their view) sufficiently guaranteed by the system of allocation. 65<br />

5.4. Conclusions<br />

Italian regional and local governments prefer a long standing<br />

system of negotiated and discretional transfers over a formula<br />

driven one. The reintroduction of local tax autonomy has made<br />

equalisation of resources more impelling than an adjustment to<br />

expenditure needs.<br />

The new constitutional framework calls for an assessment of<br />

expenditure needs only for socially strategic sectors, where<br />

essential levels of service delivery have been mandated. Essential<br />

levels are currently interpreted as implying strong uniformity in<br />

the provision of these services across the country.<br />

The new constitutional provisions also open the way for protracted<br />

bargaining on the identification of the set of services to which<br />

essential levels will have to be applied. New conflicts between the<br />

rich and the poor regions and between regional and local<br />

governments over this set of services are to be expected.<br />

Decreasing attention to need factors is partly justified by the<br />

increase in own revenue that expands the necessity of revenue<br />

equalisation. However, new methods and procedures for assessing<br />

needs related to the “essential level of services” mandated by the<br />

new Constitution will have to be devised.<br />

65<br />

It is worth noting that the absolute negative impact for the less favoured region<br />

(Campania, the largest southern region) amounted to a mere .5 per cent of its<br />

total revenues. A new agreement reached in 2005 postponed the end of the<br />

phasing out period from 2013 to 2067 but the date has not yet been implemented.<br />

148


References<br />

Giarda P., L’esperienza italiana di federalismo fiscale. Una rivisitazione del<br />

decreto legislativo 56/2000, Il Mulino, Bologna, 2005.<br />

Ministero dell’ Interno, Ricerca sui parametri obiettivi per il riparto dei<br />

trasferimenti erariali agli enti locali, Roma, 1995.<br />

Ministero dell’ Interno- Osservatorio per la finanza e la contabilità degli enti<br />

locali, Revisione del sistema dei trasferimenti erariali agli enti locali. II<br />

relazione finale, Roma, 2002.<br />

149


Appendix: Criteria for revenue sharing of VAT for Ordinary<br />

Statute Regions. 2001<br />

Criteria for revenue sharing of VAT for Ordinary Statute Regions:<br />

2001<br />

As mentioned in the text, Decree N.56/2000 established a new<br />

revenue sharing arrangement for ordinary regions, based on the<br />

personal income tax, the petrol tax and VAT. While the first two<br />

taxes were to be shared on a derivation basis, the distribution of<br />

VAT was tied to a distinct equalisation mechanism, which is<br />

illustrated as follows.<br />

The first step is the distribution of VAT among Regions according<br />

to shares of private consumption.<br />

The second step is the calculation of the regional allocations (s.c.<br />

“historical component”) which would allow each Region to balance<br />

its budget after the elimination of the previous transfers. These<br />

allocations represented the initial distribution in 2001. From 2002,<br />

they were to be gradually reduced (5 per cent a year in the first two<br />

years, 9 per cent a year in the subsequent ten years) to make room<br />

for equalisation. The allocation of this (growing) share is made<br />

according the following formula:<br />

Comp it VAT= P it + P it β∑jτ t j(xj t -xj it ) + P it (s it - s t ) + P it γ t (e it -e t )<br />

∑iP it VAT t VAT t VAT t<br />

where:<br />

i= Regions (i=1,2,..15)<br />

t= year (t=2001,2002,…2013)<br />

Comp it VAT= share of VAT for Region i in the year t, with 0≤<br />

Comp it VAT ≤1<br />

P it = population of region i in the year t<br />

β= 0.9, indicates the so called “solidarity coefficient” for the<br />

equalization of fiscal capacity;<br />

j= own and shared taxes which define the regional fiscal<br />

capacity<br />

τ t j= national average tax rate of revenue source j in the year t<br />

xj t = (∑i P it xj it )/ (∑i P it ) = average per- capita national tax base<br />

xj it = per capita tax base of region i<br />

150


s it =<br />

s t =<br />

γ t =<br />

per- capita health expenditure needs of region i<br />

(∑i P it s it )/ (∑i P it )= national average per capita health<br />

expendi-ture needs<br />

0.7, indicates the level of equalization that must be obtained<br />

with this criterion<br />

e it = standardized per capita non-health expenditure of region i,<br />

calculated through a statistical regression model 66 which<br />

assesses the extra costs due to the small dimensions of<br />

regions<br />

e t = national average standardized per capita non-health expenditure<br />

VAT t = total VAT shared among Regions in the year t.<br />

The first term of the formula determines an equal per capita<br />

distribution. This is corrected to take into account the differences<br />

in tax bases and the differences in health and total non-health<br />

expenditure needs.<br />

The second term defines a regional per-capita fiscal capacity<br />

measure and relates it to an equalisation yardstick represented by<br />

the national average per-capita fiscal capacity. The Regions whose<br />

fiscal capacity exceeds the yardstick will have the 90 per cent of<br />

their difference reduced; the opposite goes for the Regions whose<br />

fiscal capacity fall short of the yardstick.<br />

The third term is based on the difference between the health<br />

expenditure needs of each Region and the national average<br />

determined by the Health Ministry using economic and<br />

demographic indicators. Basically, it attributes different weights to<br />

the various age groups of the population with reference to different<br />

services, such as preventive care, hospitals, etc. (see table 4 in the<br />

text).<br />

Finally, the fourth term seeks to compensate for the absence of<br />

scale economies with reference to the aggregate non- health<br />

expenditures in the smallest Regions. In fact, consolidation of<br />

Regions does not appear as feasible in the present political context.<br />

66<br />

Based on a loglinear relation: e it =a+b*log(P it ).<br />

151


Table A1. The new revenue sharing of VAT for Ordinary statute<br />

Regions: criteria of distribution<br />

Regions Private<br />

Consumption<br />

% Historical<br />

<strong>Expenditure</strong><br />

% Equalization<br />

component<br />

t VAT<br />

%<br />

f=b +<br />

c+ d + e<br />

Piemonte 3335 9.3 2467 6.9 2593 7.21<br />

Lombardia 7543 21.0 281 0.8 805 2.24<br />

Veneto 3669 10.2 2125 5.9 2075 5.77<br />

Liguria 1420 3.9 1935 5.4 1978 5.5<br />

Emilia-<br />

Romagna<br />

3513 9.8 1833 5.1 1845 5.13<br />

Toscana 2803 7.8 2711 7.5 2751 7.65<br />

Marche 1179 3.3 1118 3.1 1410 3.92<br />

Umbria 603 1.7 1005 2.8 1025 2.85<br />

Lazio 3858 10.7 2661 7.4 2478 6.89<br />

Abruzzo 867 2.4 1609 4.5 1640 4.56<br />

Molise 192 0.5 544 1.5 561 1.56<br />

Campania 3106 8.6 7959 22.1 7386 20.54<br />

Basilicata 312 0.9 944 2.6 949 2.64<br />

Puglia 2373 6.6 5332 14.8 5358 14.9<br />

Calabria 1183 3.3 3432 9.5 3107 8.64<br />

Total 35958 100 35958 100 35958 100<br />

Criteria of equalisation<br />

Regions Population (b) Health<br />

expendit.<br />

<strong>Needs</strong> (c)<br />

Non-health<br />

Exp. needs<br />

(d)<br />

Revenue capacity (d)<br />

Piemonte 8.82 1.11 -0.07 -2.65<br />

Lombardia 18.57 -0.19 -2.18 -13.96<br />

Veneto 9.23 -0.11 -0.15 -3.2<br />

Liguria 3.36 1.14 0.54 0.46<br />

Emilia-<br />

Romagna<br />

8.14 1.5 0.05 -4.56<br />

Toscana 7.26 1.21 0.19 -1.01<br />

Marche 2.99 0.24 0.55 0.14<br />

Umbria 1.71 0.27 0.48 0.39<br />

Lazio 10.81 -0.7 -0.47 -2.75<br />

Abruzzo 2.63 0.08 0.54 1.31<br />

Molise 0.68 0.06 0.3 0.52<br />

Campania 11.91 -2.36 -0.72 11.71<br />

Basilicata 1.25 -0.08 0.42 1.05<br />

Puglia 8.4 -1.62 0.01 8.11<br />

Calabria 4.25 -0.56 0.51 4.44<br />

Total 100 100 100 100<br />

Source: Ministry of Treasury.<br />

152


Yet, to avoid excessive incentives to inefficiency, this equalisation<br />

component is reduced to 70 per cent with the parameter γ t .<br />

Table A1 displays the resulting allocations. The first column shows<br />

the distribution based on consumption, the second the “historical<br />

component”, and the third the distribution made according to the<br />

formula. A clear pattern emerges in favour of the less developed<br />

regions.<br />

Table A2. Share of own revenues at the different levels of<br />

government<br />

1990 1995 2006<br />

State 92 88 77<br />

Regions 2 3 13<br />

Health authorities (*) 1 1 1<br />

Provinces & Communes 5 8 9<br />

Total 100 100 100<br />

Source: Ministry of the Economy, Country General Economic Report, various<br />

years; cash revenues net of transfers. Revenues of <strong>Social</strong> Security Funds are not<br />

included.(*) In Italy health services are provided by special purpose autonomous<br />

bodies, the <strong>Local</strong> Health Authorities (Aziende Sanitarie <strong>Local</strong>i), financed by the<br />

Regions and not dependent on local governments.<br />

Table A3. Share of expenditures at the different levels of<br />

government<br />

1990 1995 2006<br />

State 63 60 54<br />

Regions 7 7 9<br />

Health authorities 16 16 17<br />

Provinces & Communes 14 17 20<br />

Total 100 100 100<br />

Source: Ministry of the Economy, Country General Economic Report, various<br />

years. °Consolidated data: transfers from one level of government to the others are<br />

included in the expenditure of the recipients. Cash expenditures net of interests<br />

paid by the state. <strong>Expenditure</strong>s of <strong>Social</strong> Security Funds are not included.<br />

153


Table A4. Simple correlation between per capita GDP and Regions’<br />

per capita expenditure 2005<br />

<strong>Expenditure</strong>s<br />

per capita GDP<br />

Recurrent expenditure 0.369767625<br />

capital expenditure -0.489654792<br />

total expenditure -0.390389736<br />

health expenditure 0.191705961<br />

total net of health -0.521283994<br />

Table A5. Per capita GDP and per capita expenditures of Ordinary<br />

statute regions. 2005<br />

Regions Populatio<br />

n<br />

Census<br />

per<br />

capita<br />

GDP<br />

recurre<br />

nt<br />

expendi<br />

capital<br />

expendi<br />

ture<br />

total<br />

expendi<br />

ture<br />

health<br />

expendit<br />

ure<br />

total<br />

net of<br />

health<br />

2001<br />

ture<br />

Piemonte 4214677 26333 2054 338 2392 1661 731<br />

Lombardia 9032554 31406 1828 236 2064 1573 491<br />

Veneto 4527694 27831 1866 287 2153 1609 544<br />

Liguria 1571783 24831 2208 296 2504 1836 668<br />

Emilia-<br />

Romagna<br />

3983346 29289 1886 225 2111 1699 412<br />

Toscana 3497806 26280 1962 288 2250 1647 603<br />

Umbria 825826 22830 2097 580 2677 1629 1048<br />

Marche 1470581 24154 1827 371 2198 1544 654<br />

Lazio 5112413 29398 2003 442 2445 1918 527<br />

Abruzzo 1262392 19621 1851 657 2508 1729 779<br />

Molise 320601 17542 1920 2280 4200 2033 2167<br />

Campania 5701931 15492 1717 518 2235 1669 566<br />

Puglia 4020707 15919 1798 347 2145 1514 631<br />

Basilicata 597768 16885 1885 1108 2993 1505 1488<br />

Calabria 2011466 15649 1824 503 2327 1423 904<br />

Italy 48151545 24182 1887 376 2263 1648 615<br />

Source: Audit Commission (Corte dei Conti), 2007, budget appropriations.<br />

154


Figure A1. The impact of the new equalisation system for the<br />

regions after two years (2002- 2003)<br />

Net balance<br />

120<br />

100<br />

80<br />

60<br />

40<br />

20<br />

0<br />

1,00<br />

0,50<br />

0,00<br />

-0,50<br />

-1,00<br />

Net balance in % of total revenues<br />

net balance 2003<br />

net balance in % total revenues<br />

-20<br />

-1,50<br />

-40<br />

-60<br />

Piemonte<br />

Source: Audit Commission<br />

Lombardia<br />

Veneto<br />

Liguria<br />

Emilia Romagna<br />

Toscana<br />

Umbria<br />

Marche<br />

Lazio<br />

Abruzzo<br />

Ordinary statute Regions<br />

Molise<br />

Campania<br />

Puglia<br />

Basilicata<br />

Calabria<br />

-2,00<br />

155


Chapter 6<br />

Equalisation of the settlement pattern and<br />

size of the municipality in Norway<br />

Grete Lilleschulstad<br />

Summary<br />

Municipalities are fully compensated for objective differences in<br />

their need for involuntary costs through the General Purpose<br />

Grant Scheme. The settlement pattern and the geographical<br />

differences in Norway challenge the perspective of equalisation of<br />

expenditures. This paper gives an overview of how these objective<br />

costs due to the settlement pattern and the size of the municipality<br />

are treated in the equalisation of expenses today. Much effort is<br />

put into developing criteria that will capture the costs due to the<br />

settlement pattern. This paper will discuss these questions, and<br />

also the method applied to determine the weights that these<br />

criteria should have in the equalisation of expenditure (the<br />

calculation matrix).<br />

In addition, the vertical grants for the remote areas and Northern<br />

Norway are accounted for. These grants are based on political<br />

priorities.<br />

In May 2008, the <strong>Government</strong> aims to present a white paper on a<br />

revised General Purpose Grant Scheme, including a revised<br />

calculation matrix.<br />

6.1. Why make allowances to the settlement pattern and the<br />

size of the municipality in the Grant scheme<br />

The Norwegian geography and settlement pattern lead to special<br />

challenges in the equalisation of expenditures. The country is<br />

thinly populated, with sparsely inhabited municipalities and long<br />

distances to the municipality centre. There are a total of 430<br />

156


municipalities. The smallest municipality is the islands of Utsira,<br />

with approximately 200 inhabitants. The largest Norwegian<br />

municipality is the capital, Oslo, with almost 0.6 million<br />

inhabitants.<br />

Box 1. Facts about Norway<br />

There are 430 municipalities. The areas and the population density of the<br />

various municipalities vary. There are municipalities with a large area and a<br />

dispersed population, and there are highly urbanised municipalities which are<br />

densely populated.<br />

The populations vary from about 200 to almost 600 000 inhabitants, and ¼ of the<br />

whole population of 4.7 million live in the capital-region.<br />

Next to Iceland, Norway has the lowest population density in Europe, with 15<br />

inhabitants per square kilometre.<br />

44.4 per cent of the area is mountains and plateaus, while 38.2 per cent is<br />

woodland.<br />

Coastline:<br />

Mainland without fjords and bays: 2 532 km<br />

Mainland with fjords and bays: 25 148 km<br />

Islands: 83 281 km<br />

Municipal and county finances:<br />

Nearly 70 per cent of the revenues go through the General Purpose Grant<br />

Scheme.<br />

Part equalisation of local taxes and full compensation for involuntary production<br />

costs.<br />

Composition of revenues:<br />

• <strong>Local</strong> taxes: 49 per cent<br />

• General grant: 19 per cent<br />

• Earmarked grants: 10 per cent<br />

• Fees and charges: 15 per cent<br />

• Other revenues: 7 per cent<br />

The equalisation mechanisms in the General Purpose Grant<br />

Scheme give full compensation for involuntary production costs of<br />

national welfare services such as primary education, health, social<br />

welfare (financial support for welfare clients) and care of the<br />

elderly. Settlement patterns and sizes of the municipalities are<br />

criteria in the distribution formula to secure a fair distribution,<br />

and to make small and sparsely inhabited municipalities capable of<br />

providing equivalent public services.<br />

157


In addition, vertical grants are in place in order to obtain regional<br />

policy goals, and not from strictly objective criteria. These are<br />

specific grants to municipalities in Northern Norway and a grant<br />

for remote and small municipalities. This extra contribution is<br />

supposed to ensure good living conditions in provincial<br />

municipalities (cf. part 4).<br />

Figure 1. Municipalities and population, Norway<br />

Inhabitants<br />

6.2. Settlement pattern – criteria<br />

Some municipalities have involuntary high costs for primary and<br />

secondary schools because of the settlement pattern. Due to<br />

limitations in the accepted travelling time that pupils can have,<br />

there is a decentralised school structure in sparsely inhabited<br />

municipalities. Consequently, there will be small schools with a<br />

relatively high number of teachers per pupil in rural areas,<br />

158


whereas in urban areas the size of the school will limit the number<br />

of pupils. Norwegian primary and secondary schools rarely have<br />

more than 400 students. The calculation matrix for primary and<br />

secondary education is composed of several criteria for settlement<br />

patterns which represent involuntary costs caused by these<br />

conditions. Much effort is put into developing a reliable calculation<br />

matrix. The extensive analysis is based on a division of<br />

municipalities into smaller geographical unities (basic statistical<br />

unit and zone) and on information about distances from Norwegian<br />

road databases. The estimations and the annual update are<br />

performed by Statistics Norway.<br />

About basic statistical unit and zone<br />

Basic statistical unit<br />

Each municipality is divided into smaller geographical units, socalled<br />

”basic statistical units”. The division is undertaken by<br />

Statistics Norway according to a set of determined criteria – i.e. as<br />

homogeneous units as possible regardless of municipality. In total,<br />

there are about 15 000 basic statistical units. In every unit a<br />

digital centre has been established. The premise for the further<br />

analysis is that the population lives in this centre. Further on,<br />

every distance to and fro is measured according to this digital<br />

centre. The municipalities also use this material for long-term<br />

planning purposes (area plans, municipality plans etc.).<br />

Zone<br />

The basic statistical units are aggregated into zones, i.e. a<br />

geographically connected area of basic statistical units. One zone<br />

should have minimum 2 000 inhabitants, and in each zone a centre<br />

is established. On average, 2 000 inhabitants enable a mediumsized<br />

elementary school according to Norwegian standards.<br />

Therefore, small municipalities (i.e. less than 2 000 inhabitants)<br />

only have one zone. The zones are not made based on actual service<br />

provision, but on geography, population and municipality<br />

boundaries. Because of the requirement of a minimum number of<br />

inhabitants within the zone, sparsely and densely populated<br />

municipalities have very different zonations. Sparsely inhabited<br />

municipalities will have zones with large rural areas, and thus<br />

159


long travelling distances. Densely populated municipalities will<br />

have zones with small rural areas and short travelling distances.<br />

Based on the databases of zones and basic statistical units, there<br />

are three criteria:<br />

1. “estimated travel time”,<br />

2. “zone”,<br />

3. “neighbour”.<br />

These are included in the calculation matrix for primary and<br />

secondary education. The travel time-criterion is estimated in<br />

minutes, while the two estimates of distance are in kilometres. If<br />

possible, the distances are estimated along a road (if the distance<br />

from the centre in the basic statistical unit to a road is short, the<br />

distance will be inscribed to the nearest road). In the case of<br />

distances across water, it is considered whether or not there is a<br />

ferry connection.<br />

Criterion 1 – Estimated travel time<br />

The estimation of this criterion uses the division into basic<br />

statistical units and measures the shortest distance from the<br />

digitally determined centre in every basic statistical unit to the<br />

municipal centre, i.e. the basic statistical unit where the<br />

administration centre in the municipality is situated.<br />

This criterion differs between various types of rural municipalities.<br />

Sparsely inhabited municipalities with large rural areas or<br />

municipalities with the population scattered on several islands are<br />

favoured. These municipalities generally have an extensive area<br />

with longer travel time to the municipal centre than the national<br />

average. Also the larger cities generally cover large geographical<br />

areas. Therefore, this criterion is not sufficient to distinguish<br />

between the large cities and the typical rural municipalities, i.e.<br />

the large municipalities will be favoured.<br />

Except for the smallest ones, it is not realistic for the<br />

municipalities to organise the school only in the municipality<br />

centre. Most municipalities need a decentralised school structure.<br />

This is taken into account in the estimation of involuntary<br />

production costs in the two other criteria (zone and neighbour).<br />

160


These measures are based on a division of the municipalities into<br />

zones.<br />

Criterion 2 – Travel distance within the zone<br />

The criterion “travel distance within the zone” (“zone”), is the sum<br />

of the distance from the centre of the inhabitants’ own basic<br />

statistical unit to the centre of the zone. The centre of the zone is<br />

defined as the basic statistical unit with the highest number of<br />

inhabitants. If there are several zones within a municipality, the<br />

estimated distances for the different zones are aggregated. This<br />

criterion is based on the same principles as the criterion<br />

“estimated travel time”.<br />

Criterion 3 – Travel distance to nearest neighbour - basic statistical<br />

unit<br />

The criterion “travel distance to nearest neighbour - basic<br />

statistical unit” (“neighbour”) is the inhabitants’ distance from the<br />

centre within their own basic unit to the centre in the nearest basic<br />

statistical unit (within the same zone), aggregated for all the<br />

inhabitants in the municipality.<br />

Through the criterion “zone”, the municipalities’ higher costs due<br />

to the settlement pattern is taken into consideration, based on<br />

economy of scale by operating with schools that are average size or<br />

large.<br />

The “neighbour-criterion” is composed in order to ensure a more<br />

nuanced picture of the settlement in the municipality, and the<br />

possible need to decentralise the schools within the zone. If the<br />

basic statistic units are extensive, the distances between the basic<br />

statistic units will be correspondingly large. Therefore the<br />

distances can present an obstacle for the municipality to utilise<br />

prospective economies of scale. More schools have to be established<br />

in order to give the municipality a natural school structure.<br />

Statistical analyses show that the criteria are relatively strongly<br />

correlated, but together these criteria give a good impression of the<br />

settlement in the country and take into consideration the different<br />

levels of costs of the primary school services.<br />

161


6.3. From criteria to weights in the calculation matrix<br />

The calculation matrix is used to estimate involuntary production<br />

costs. Municipalities with estimated expenditure needs above the<br />

national average are fully compensated. Municipalities with<br />

estimated expenditure needs below the average have to contribute<br />

to the financing. In the total calculation matrix for the<br />

municipalities, and in the partial calculation matrix for primary<br />

and secondary education, which is a part of the total calculation<br />

matrix, the criteria will have the weights shown in table 1.<br />

Table 1. Weight of the settlement criteria in the calculation matrix<br />

for primary school and in the total calculation matrix for the<br />

municipalities in the General Purpose Grant Scheme<br />

Criteria<br />

Weight in the total<br />

calculation matrix<br />

for the municipalities<br />

(per cent)<br />

Weight in the calculation<br />

matrix for primary and<br />

secondary school (per<br />

cent)<br />

Travelling time 1.5 4.8<br />

Zone 1.0 3.0<br />

Neighbour 1.1 3.4<br />

Sum 3.6 11.2<br />

Total sum 100.0 100.0<br />

Table 1 shows that costs due to settlement patterns are presumed<br />

to amount to 3.6 per cent of the municipalities’ total expenditures<br />

on primary and secondary education, health and care services and<br />

the administration sector.<br />

The weights in the calculation matrix are determined through<br />

regression analysis. The regression analysis includes factors that<br />

may influence expenditures. The dependent variable is gross<br />

expenditures, and explanatory variables are the number of<br />

inhabitants divided according to age and socio-economic variables<br />

(number of divorced or separated, the death rate etc.).<br />

162


The analysis of primary and secondary education is different from<br />

the analysis of other sectors. The dependent variable is standard<br />

number of teaching hours (not gross expenditures). The number of<br />

teaching hours is determined according to standards based on the<br />

size of the school, class size, teaching hours a week etc. The<br />

variation in hours is analysed by simple multiple regression<br />

analysis (I).<br />

(I) Yi = a + bxi + czi + ei<br />

In addition to the settlement pattern criteria, the number of 6-15-<br />

year old inhabitants is also included in the analysis. All these<br />

variables are significant and are used in the further estimations.<br />

To set the weights of the respective criteria in the total calculation<br />

matrix, each variables’ part of the total variation in the dependent<br />

variable (Y) is estimated subsequently.<br />

6.4. Size of the municipality – compensation for diseconomies<br />

of scale<br />

The basis criterion<br />

The size of the municipality (number of inhabitants) is of<br />

significant importance for the costs of municipal service provision.<br />

The existence of diseconomies of scale in municipal service<br />

production is well documented. Municipalities with few<br />

inhabitants therefore need a higher level of income per capita than<br />

municipalities with a larger number of inhabitants in order to<br />

provide equal standards of the public services.<br />

Generally, the settlement pattern criteria benefit the small<br />

municipalities since they are often, but not always, sparsely<br />

inhabited. But the size of the municipality is explicitly considered<br />

in two different ways in the Norwegian system. First, diseconomies<br />

of scale are considered through objective criteria in the<br />

equalisation of expenditures with a scale criterion (the basis<br />

criterion). Furthermore, there is a special grant for small<br />

municipalities and grants for municipalities in Northern Norway;<br />

these grants are given based on regional policy aims.<br />

Independently of settlement pattern, the municipalities are<br />

compensated for diseconomies of scale within administration,<br />

163


agricultural issues, environmental issues and primary and<br />

secondary education through the basic criterion. It is presumed<br />

that municipalities with few inhabitants need a higher level of<br />

income to provide the same services.<br />

Criteria can be shaped in several ways. In the Norwegian system,<br />

the criterion is designed in such a way that every municipality gets<br />

the same amount regardless of its size (fixed sum grant). This<br />

means that small municipalities get a high amount per capita, and<br />

large municipalities get a small amount per capita. Totally the<br />

weight is set to 2.5 per cent in the total calculation matrix, i.e. it is<br />

expected that 2.5 per cent of the local governments’ total<br />

expenditures on national welfare services are related to<br />

diseconomies of scale. The weight is determined through an<br />

objective analysis. The basic criterion is thus an important<br />

contribution to providing equal standards of service provision<br />

regardless of the size of the municipalities.<br />

Grants for remote regions<br />

A main political goal is to keep the settlement pattern and the local<br />

structure in rural Norway. Like most countries, Norway<br />

experiences centralisation and pressure on the urban areas. In<br />

addition to the objective criteria in the calculation matrix (the<br />

settlement pattern criteria and the basic criterion), Norway has<br />

the grants to municipalities in remote regions and Northern<br />

Norway. Combined, these mechanisms are intended to enable<br />

sparsely inhabited municipalities, municipalities in remote regions<br />

and in Northern Norway to provide services that are equal to, and<br />

partly better than, those provided by municipalities in more<br />

central areas.<br />

Municipalities with less than 3200 inhabitants and a tax income<br />

per capita below 110 per cent of the average receive the grant for<br />

remote regions. This grant is a fixed amount per municipality.<br />

The grant to Northern Norway is a special regional political<br />

instrument. The objective for the Northern Norway grant, which is<br />

granted as a flat per capita amount, is to make the municipalities<br />

and counties in Northern Norway able to provide the same or a<br />

higher level of municipal services compared to the rest of the<br />

164


country. This will also arrange for industrial and commercial<br />

development in the area.<br />

6.5. Current work<br />

The General Purpose Grant Scheme, including the calculation<br />

matrix, for the Norwegian system is now being revised. The<br />

<strong>Government</strong> aims to present a white paper on the revised General<br />

Purpose Grant Scheme to the Storting (the Norwegian Parliament)<br />

in May 2008. The report from “the Borge committee” ((NOU 2005:<br />

18) Fordeling, forenkling, forbedring) is part of the foundation of<br />

the work with the new calculation matrix. The Borge committee<br />

recommended analysing the expenditures for primary and<br />

secondary schools on the basis of actual gross operating<br />

expenditures, instead of on standard number of teaching hours. In<br />

recent years, the municipalities have gained more freedom and<br />

room for local manoeuvres in how to organise primary and<br />

secondary schooling. Further, the Borge committee recommended<br />

including the settlement pattern criteria in the partial calculation<br />

matrix for health and care services.<br />

165


Chapter 7<br />

<strong>Measuring</strong> expenditure needs: Japan’s<br />

experiences<br />

Nobuki Mochida<br />

7.1. Decentralisation in Japan<br />

The Japanese model<br />

In Japan’s overall fiscal system, while the majority of expenditures<br />

are done at the local level, only very limited autonomy is available<br />

to local governments in their spending decisions. Also on the<br />

revenue side decentralisation is limited, and the authority to<br />

determine tax base and rates lies with the central government. The<br />

vertical financial imbalance can be seen in Table 1, which presents<br />

the relative share of all tax revenues and expenditures of the two<br />

levels of government. As suggested by the table, the central<br />

government has collected more than 60 per cent of the total tax<br />

revenues since 1970, while it has expended less than 45 per cent of<br />

the total tax revenue spent.<br />

Table 1. Vertical Fiscal Imbalance (in per cent)<br />

Tax revenue received by Tax revenue received by<br />

Central <strong>Local</strong> Central <strong>Local</strong><br />

1970<br />

1980<br />

1990<br />

2000<br />

2005<br />

67.5<br />

64.1<br />

65.2<br />

59.7<br />

60.0<br />

32.5<br />

35.9<br />

34.8<br />

40.3<br />

40.0<br />

50.8<br />

46.0<br />

47.0<br />

43.0<br />

40.6<br />

49.5<br />

54.0<br />

53.0<br />

57.0<br />

59.4<br />

(Source) Ministry of Internal Affairs and Communication (2005)<br />

Mochida and Lotz (1999) compare the decentralisation of the public<br />

sector in Japan and the Northern European countries as follows.<br />

166


• Central-local relations are regulated by law. But legislation<br />

in Japan allows the central government to control local<br />

expenditure decisions on a discretionary basis through a<br />

large number of “delegated functions” financed by specific<br />

purpose grants. The local expenditure policies of the Nordic<br />

countries are now controlled by legislation; no central civil<br />

servant can overrule a local politician if he keeps his<br />

activities within the law.<br />

• The Nordic countries employ a piggyback type of local<br />

income tax for which each local authority determines its<br />

own tax rate. In Japan, local tax revenues are mainly<br />

derived from a variety of own income, property and<br />

consumption with very limited authority to vary tax rates.<br />

This way the Japanese system combines a Continental style<br />

tax-sharing type of financing with a Northern European<br />

decentralisation of expenditures. There are signs that this<br />

type of financing is not suitable when expenditures are so<br />

highly decentralised.<br />

Box 1. Some explanations why Japan wants to decentralise public<br />

sectors<br />

• There is an increasing concern over inefficient spending decisions made<br />

by the central government and the extensive regulations and uniform<br />

standards it imposes on local representatives. The “one-size-fits-all”<br />

approach has failed.<br />

• Faced with a gross debt amounting to 150 per cent of the GDP,<br />

revamping fiscal relations across levels of government is of paramount<br />

importance. It requires cutting fiscal transfers<br />

• And improving the cost-effectiveness of local public spending.<br />

• Since outlays on healthcare and other welfare programmes are largely<br />

under local government responsibility, population aging will change<br />

the demand pattern for local public services, across both spending<br />

categories and jurisdictions.<br />

• A new ideology of neo-liberalism has emerged among policy-making<br />

elites. As a result, the traditional policy goal of “promotion of balanced<br />

regional development with rural areas” has been increasingly eroded.<br />

Source: Machida (2004) Fiscal Decentralization and State-<strong>Local</strong><br />

Finance: Fundamental Perspectives<br />

167


The central government defends its use of controls by noting that it<br />

helps standardise welfare and health programmes, which ought to<br />

be uniform across the country, and further claims that without<br />

national government involvement, economically fragile localities<br />

may fail to provide adequate services for their residents. The<br />

central government also believes that equalisation of different<br />

social services should take precedence over local autonomy and<br />

decentralisation. Over the last few years, however, the public has<br />

increasingly called for deregulation and decentralisation of Japan’s<br />

centralised intergovernmental relations. Public statements claim<br />

that Japan’s centralised system of government will eventually be<br />

unable to resolve issues generated by the economic globalisation<br />

and the fiscal stress caused by post-bubble economic blues (see box<br />

1 above).<br />

Drive to decentralise<br />

Reflecting these general outcries, the government finally enacted<br />

The Law for the Promotion of Decentralization, on May 15, 1995<br />

and set up a committee, headed by a prominent businessman, to<br />

study different methods of decentralisation. After intensive debate<br />

and investigation, the committee submitted its final conclusion in<br />

1998. The report suggested a virtual elimination of agencydelegated<br />

functions, and proposed to term a large portion of<br />

prefectural and municipal affairs as ’<strong>Local</strong> <strong>Government</strong> Services.’<br />

In 1999, to implement the new pattern of central-local<br />

relationships, the national legislature eradicated 432 of the laws<br />

that formed the basis for agency-delegated functions.<br />

In addition to the deregulation of administrative control,<br />

intergovernmental financial relations have also been reformed<br />

recently in order to strengthen local own revenues and to reduce<br />

the heavy reliance on the central government coffer. In FY 2002,<br />

Prime Minister Koizumi launched an ambitious reform of the three<br />

main components of local government financial resources -<br />

earmarked grants, local taxes and the local allocation tax – the so<br />

called ”Trinity Reform”.<br />

• A cut in earmarked grants to local governments by around 4<br />

trillion yen over the period FY 2004-2006.<br />

168


• A streamlining of the <strong>Local</strong> Allocation Tax. The <strong>Local</strong><br />

Allocation tax has been cut by 5 trillion yen, for the first<br />

time since the 1950s, over the period FY 2001-2006.<br />

• A transfer in tax resources from the central to local<br />

governments of about 3 trillion yen. The government cut the<br />

national personal income tax and increased local<br />

government reliance on the individual inhabitant tax by FY<br />

2007. In that event, standard rates for the individual<br />

inhabitant tax on income will be flattened so as to<br />

strengthen the benefit principle and avoid a distribution of<br />

tax resources across local governments that would be too<br />

uneven.<br />

As the foregoing indicates, the question of decentralisation and<br />

deregulation has developed into one of the most contentious<br />

political issues in Japan. The demand for reform strikes at some of<br />

the most critical prerogatives of the central government. The<br />

central agencies feel that deregulation and decentralisation would<br />

reduce their power, which they are reluctant to relinquish.<br />

Changing some of Japan’s most entrenched political practices,<br />

therefore, seems extremely difficult to accomplish.<br />

7.2. Intergovernmental relations<br />

The role of the Ministry of Internal Affairs and Communications<br />

The organisation of the way central control is exercised differs<br />

from that of other countries. In the Nordic countries, the system of<br />

local government control is an example of cooperation and<br />

negotiation between central and local government. The MOI<br />

(Ministry of the Interior) and the MOF represent the central<br />

government, while the local governments are represented by the<br />

<strong>Local</strong> <strong>Government</strong> Associations (Lotz 2006). In Japan it is not the<br />

local government organisations but the Ministry of Home Affairs<br />

that represents the interests of local governments as a whole. The<br />

conflicts openly displayed in Japan between the MOF and the<br />

Ministry of Home Affairs would not be acceptable for the<br />

governments of the Nordic countries, and neither would it satisfy<br />

their local authorities.<br />

Understanding this unique aspect requires paying attention to the<br />

change in the role of the ministry-level unit that oversees<br />

169


administration (Akizuki 2001:65-68). The pre-war Ministry of<br />

Interior (Naimusho), which had tremendous power covering all<br />

areas of domestic affairs, was broken up and replaced by the<br />

scaled-down Ministry of Home Affairs, or MOHA (Jichisho), in<br />

1960. In 2001, the MOHA was merged with the Ministry of Posts<br />

and Telecommunications to become the Ministry of Internal Affairs<br />

and Communications, or MIC. MIC engages in fiscal supervision of<br />

local bodies as a central government ministry, while at the same<br />

time representing the interests of local governments within central<br />

government departments.<br />

The incentives for the MIC and local governments to cooperate<br />

with one another are easy to understand. First, in all problems<br />

related to local governments—such as a reduction of the local<br />

allocation tax— the MIC protects the interests of local<br />

governments and engages in tough fights with other central<br />

government ministries (especially the Ministry of Finance) that<br />

want to get their hands on local government finances. <strong>Local</strong><br />

governments, of course, want to maintain this mechanism. And,<br />

from the MIC’s perspective, if it loses the support of local<br />

governments—which forms the main foundation of its authority—<br />

it cannot survive as a significant player in central bureaucracy.<br />

The relationship between the MIC and local governments has<br />

become more and more interdependent with the passage of time.<br />

Structural reform<br />

While the key strength of Japan’s “controlled decentralisation” is<br />

its stability in the constitutional, legal, financial and<br />

administrative framework, one of the concerns about it is the lack<br />

of flexibility (Akizuki 2001:79-84). Put another way, if politicians<br />

and bureaucrats have vital interests in the local finance system, it<br />

is tremendously difficult to change the status quo: (1) the Ministry<br />

of Internal Affairs and Telecommunications (MIC), responsible for<br />

<strong>Local</strong> Allocation Tax, has opposed LAT reform; (2) Spending<br />

ministries have resisted all attempts to reduce conditional grants;<br />

and (3) the Ministry of Finance (MOF) opposed a transfer of tax<br />

revenue source from central to local government. Consequently,<br />

the local government finance system has been based on a subtle<br />

political equilibrium among the main political actors (Kitamura<br />

2006:16).<br />

170


How can we explain why Prime Minister Koizumi could implement<br />

such a radical “Trinity” of local finance reforms Political scientists<br />

argue that both the introduction in 1994 of an electoral system<br />

that combined single-member districts and proportional<br />

representation, and the reorganisation of the central ministries<br />

and strengthening of the position of the prime minister and cabinet<br />

(implemented in 2001), have weakened the coalition of LDP<br />

politicians, bureaucrats, and interest groups that dominated postwar<br />

Japan (Kitamura 2006). Turning our attention to local finance<br />

reforms under the Koizumi administration, two points are worth<br />

mentioning in this respect.<br />

• The Council of Economic and Fiscal Policy (CEFP) was to be<br />

established to investigate all government policies from a<br />

financial perspective. The Koizumi administration succeeded<br />

in setting the macro targets at the CEFP, thereby forcing the<br />

MIC to reluctantly accept the simplification of the <strong>Local</strong><br />

Allocation Tax system. The amount of the LAT was curtailed<br />

for the first time in fifty years.<br />

• A tentative Consultative Council between central ministries<br />

and local representatives was to be set up to come up with the<br />

plan for the size of cuts in specific purpose grants and transfer<br />

of tax resources. After painstaking debates and tough<br />

negotiations, both sides have successfully agreed upon the<br />

transfer of tax resources of 3 trillion yen via cuts in grants of 4<br />

trillion yen.<br />

7.3. Fiscal equalisation<br />

Arguments for equalisation<br />

Japan needs strong equalisation systems for two reasons.<br />

Probably, the strongest case for equalisation transfer has been<br />

based on the premise of “horizontal equity among the citizens”<br />

(Buchanan 1950). Equity requires that the equalisation should be<br />

full, and this appears to be the logic behind Japan’s <strong>Local</strong><br />

Allocation Tax. As pointed out by the Recommendation by Prof.<br />

Carl S. Shoup of 1950, special support from the national<br />

government for the exceptionally poor areas is necessary, so that<br />

they are in fact, as well as in principle, free to vary the amount of<br />

services they supply to their citizens, through changes in their own<br />

171


tax rates (General Headquarters Supreme Commander for Allied<br />

Power 1949). Boadway, Hobson and Mochida (2001) also<br />

characterises local residential taxes as being proportional to<br />

income and local expenditure as being distributed on a slightly<br />

progressive basis in Japan. The progressivity of local budgets leads<br />

to differences in net fiscal benefits and calls for full or more than<br />

full equalisation of residence-based taxes in Japan.<br />

Another case for Japan’s equalisation is a guarantee of adequate<br />

revenue to local governments to allow them to provide a national<br />

standard level of public services. Its focus is not on making sure<br />

that all local governments have the fiscal capacity to deliver<br />

reasonably comparable service to their residents at reasonably<br />

comparable levels of taxation like in Canada. Instead, it is<br />

designed to ensure that common standards in quality or outcomes<br />

in public services are achieved. Put differently, the <strong>Local</strong> Allocation<br />

Tax is uniquely designed to reflect the Japanese reality where the<br />

law allows the central government to control local expenditure<br />

decisions on a discretionary basis through a large number of<br />

“delegated functions” financed by specific purpose grants. During<br />

the 1950s and 1960s, Japan was an economically poor country.<br />

Depending on where one lived, there were significant gaps in<br />

schools and other educational facilities as well as in roads,<br />

hospitals, waterworks, and medical care for elderly people. The<br />

Japanese public strongly desired public services that conformed to<br />

uniform standards.<br />

Two models: grants or a solidarity model of equalisation<br />

Horizontal imbalances are reduced by equalisation arrangements.<br />

These can be designed by applying one of two very different forms<br />

of techniques; we shall call them the grants method and the<br />

solidarity method (Mochida and Lotz 1999). The most common<br />

model by far is a central government grant to the poor local<br />

authorities, in order to bring their revenue up to a certain national<br />

standard like in Japan. The other method, the solidarity method –<br />

sometimes called the “Robin Hood” model, is used in Germany<br />

(among the länder), in Denmark, and in Sweden. It removes<br />

resources from the rich authorities and gives to the poor; no central<br />

funds are needed.<br />

172


The grant model like in Japan would have some advantages over<br />

the solidarity model. It can be seen that the this model could<br />

contribute to "nation building" in countries where regional conflicts<br />

are very serious like in Canada, Korea, and Japan, while the<br />

solidarity model is likely to create conflicts between regions, as is<br />

evident from the experiences of Germany, Sweden and Denmark.<br />

However, there are, of course, serious concerns connected to the<br />

grant model. First, a lack of central funds in many countries<br />

results in insufficient equalisation reaching only the poorest<br />

authorities, which means that the more well-to-do authorities and<br />

their wealth remain outside the system. For example, the Tokyo<br />

metropolitan government whose revenue exceeds its need is<br />

neither eligible for the grants nor obligated to contribute money for<br />

fifty years. Second, under the grants model everybody seems to get<br />

something and nobody seems to pay. But this is not true because<br />

real redistribution is less visible (“fiscal illusion”) than in the<br />

solidarity model. Judging from Japan’s experiences, local<br />

politicians, civil servants and voters would take less care with<br />

money that is given as LAT.<br />

Degree of equalisation<br />

In Japan, there are some similarities in public thinking with the<br />

NCs. Equal access to public goods and fair sharing of the burden to<br />

finance these goods are viewed as essential for economic and social<br />

development. By reflecting these views, the formula means that<br />

the national government transfer will fill 100 per cent of the gap<br />

between each region’s fiscal need and its fiscal capacity. Such<br />

formulas are the basis for the whole idea of delegating<br />

responsibilities to local governments; this is the intellectual basis<br />

for the systems in Japan and the NCs as well as in countries like<br />

Australia, Korea and the United Kingdom (Ma 1996). One<br />

difference, however, is the degree of equalisation. The political<br />

hesitation to ensure full equalisation would lead to a certain<br />

acceptance of marginal variations and inequalities in some<br />

countries. In Japan, equalisation would fill the full gap, aiming at<br />

guaranteeing adequate revenue to local governments to allow them<br />

to provide uniform-but not just comparable, like in NCs or<br />

Canada, Australia-levels of public services.<br />

173


Figure 1. The degree of equalisation (per capita)<br />

400,000<br />

Yen<br />

350,000<br />

300,000<br />

250,000<br />

200,000<br />

150,000<br />

100,000<br />

50,000<br />

0<br />

tokyo<br />

kanagawa<br />

saitama<br />

chiba<br />

osaka<br />

fukuoka<br />

aichi<br />

hyogo<br />

sizuoka<br />

kyoto<br />

hirosima<br />

ibaragi<br />

gunnma<br />

miyagi<br />

nara<br />

gifu<br />

tochigi<br />

okayama<br />

mie<br />

siga<br />

okinawa<br />

kumamoto<br />

ehime<br />

nagano<br />

fukusima<br />

kagawa<br />

hokkaido<br />

yamaguchi<br />

niigat a<br />

nagasaki<br />

isikawa<br />

toyam<br />

oita<br />

wakayama<br />

akagosima<br />

yamagata<br />

miyazaki<br />

aomori<br />

saga<br />

yamanasi<br />

iwat e<br />

akit a<br />

tokusima<br />

fukui<br />

kochi<br />

tottori<br />

simane<br />

the local acclocation tax<br />

local tax revenue<br />

The degree of equalization (per capita)<br />

174


The degree of equalisation is extremely high (see figure 1). For<br />

example, the Aichi prefecture, where Toyota Motor is located, collects<br />

local taxes of 143,000 yen per capita, while Shimane, a remote area of<br />

the western part, collects only 87,000 yen. Looking at total resource<br />

from the LAT transfer and local tax, the latter’s figure increases to<br />

341,000 yen compared to the former’s 158,000 yen. Overall, some<br />

poor and/or remote jurisdictions end up having more financial<br />

resources than are available to the richest ones.<br />

In conclusion, the actual degree of equalisation was perhaps more<br />

important before the 1970s, when the transfer system contributed<br />

significantly to equality (Mochida 2001:101-104). Since then,<br />

regional fiscal disparities have been reduced, there has so to speak<br />

been less ‘inequality’ to fix through local allocation tax, and<br />

subsequently the intensity of the equalisation effect has fallen. But<br />

this does not mean that equalisation is no longer necessary;<br />

equalisation is crucial for a country with a decentralised public<br />

sector as in Japan.<br />

7.4. <strong>Expenditure</strong> needs<br />

Formula for local allocation tax<br />

<strong>Local</strong> allocation tax has continued to the present with some minor<br />

alterations. Since 1954, the framework of the local allocation tax<br />

has been founded on the former distribution tax, enforced between<br />

1940 and 1949, while retaining the formula used in the<br />

equalisation grant for the distribution of funds to localities. In the<br />

local allocation tax system, the total amount to be distributed to<br />

local authorities is a fraction of yields from major national taxes:<br />

funding pool = 0.32 × (personal tax + corporate tax) + 0.358<br />

×alcoholic tax + 0.295×VAT+ 0.25 ×tobacco tax. These prescribed<br />

percentages of five major national taxes are apportioned among<br />

local bodies in proportion to the difference between standard fiscal<br />

need and standard fiscal capacity. It is paid annually to local<br />

governments whose basic financial needs exceed their basic<br />

financial revenues. Rich localities with revenue that exceeds need<br />

are neither eligible for the grants nor required to contribute money<br />

for fiscal adjustment, as is the case in some countries (For more<br />

detail of the formula, see Mochida 2008: Chapter 6).<br />

175


What is expenditure need<br />

One of the objectives of Japan’s equalisation is to guarantee<br />

adequate revenue to local governments in order to allow them to<br />

provide a national standard level of public services across the<br />

country. This objective means taking into account not only the<br />

difference in the fiscal capacity of local governments, but also their<br />

expenditure needs. Aging and internal migration flows will affect<br />

individual local governments differently, reinforcing demands on<br />

the equalisation scheme. The share of population over the age of 65<br />

varies significantly across prefectures, from 25 per cent in<br />

Shimane to 13 per cent in Saitama. The share even reaches 40 per<br />

cent in some municipalities with a clear urban/rural divide.<br />

The concept of expenditure need encompasses two notions. The<br />

first is the relative demand for public services. A local government<br />

with a larger proportion of elderly people requiring long term care<br />

or with a higher ratio of school-age children has a significant need<br />

for primary education. This is referred to as the “workload” facing<br />

the public services. It is usually measured by the sociodemographic<br />

and economic characteristics of local residents. The<br />

second dimension of expenditure need is the financial cost of<br />

providing a given set of services. The cost difference can be due to<br />

diseconomies of scale where fixed costs must be spread among a<br />

smaller population, higher costs of serving cold areas (fuel<br />

expenses) or sparsely populated areas (wage expenses), differences<br />

in costs of promotion of regional economy. However, the difference<br />

in the efficiency and effectiveness of government service delivery<br />

should not be taken into account when we measure the<br />

expenditure need of a particular local government.<br />

The complexity of measuring expenditure need<br />

The next problem to be solved is how we should quantify<br />

expenditure need. The two schools of thought suggest different<br />

approaches to quantifying expenditure need, though in practice the<br />

concepts may be muddled and the methods mixed. The first<br />

approach points at the need for empirical work studying local<br />

authority needs as revealed in the local budgets and related to<br />

differences in some suitable objective indicators. The second<br />

approach leads to a more normative approach. The central<br />

government and Parliament decide what would be the “norm” or<br />

“standard” cost to fulfil the function of supplying a given delegated<br />

176


service that meets the criteria expected by Parliament. Japan’s<br />

fiscal equalisation system has adopted the latter approach for fifty<br />

years, reflecting a Japanese reality where the law allows the<br />

central government to control local expenditure decisions on a<br />

discretionary basis through a large number of “delegated<br />

functions” financed by specific purpose grants.<br />

Like the system developed in Australia, measuring expenditure<br />

requires the design and costing of a standard for each spending<br />

area. Before the calculation of standard expenditure needs, public<br />

services for each prefecture and municipality are divided into<br />

particular service items by function (gyôsei-kômoku). In the<br />

prefectures, there are 24 service functions, such as police, roads<br />

and bridges, and primary school; for each municipality, there are<br />

24 service functions, such as city planning, parks, and garbage<br />

collection.<br />

Standard expenditure needs are calculated as the number of<br />

measurement units by multiplying the unit cost, adjusted by<br />

modification coefficients: standard fiscal need= unit cost×<br />

measurement unit× adjustment coefficient. The total basic need in<br />

each locality is the sum of the amounts needed for all service items<br />

combined.<br />

The following provides the numerical sample of major functions:<br />

• Fire/ambulance: 10,800 yen× population (person) ×<br />

modification coefficients.<br />

• Road improvement: 380,000 yen× road length (km) ×<br />

modification coefficients.<br />

• Elementary school: 973,000 yen× No. of classes (class) ×<br />

modification coefficients.<br />

• Welfare for the elderly: 72,100 yen× population 65 years or<br />

older × modification coefficients.<br />

• Waste treatment : 6,570 yen× population (person) ×<br />

modification coefficients.<br />

• Agricultural promotion: 64,200 yen× No. farming<br />

households × modification coefficients.<br />

• Census registration: 1,670 yen× No. registered families ×<br />

modification coefficients.<br />

177


All these steps require numerous judgment calls and adjustments<br />

in order to fairly measure the expenditure needs for each local<br />

government. The first step is to select measurement units. A<br />

measurement unit or “workload” reflects the number or size of the<br />

beneficiaries of a particular expenditure. For example, a<br />

measurement unit for teachers’ salaries is the number of teachers,<br />

that of police is the number of police officers, and that of roads is<br />

the length of roads. The second step is to determine a unit cost.<br />

Unit cost is a kind of net standard cost per measurement unit for<br />

each service item. Assuming a representative local body with<br />

standard conditions and scale, the unit cost for each service item is<br />

determined by The Ministry of Internal Affairs and<br />

Telecommunications. The unit cost can be calculated by deducting<br />

earmarked revenue from gross cost and dividing by the number of<br />

measurement units.<br />

The unit cost, however, is uniform throughout the country, and no<br />

consideration is given to either the unique services or the special<br />

circumstances of localities. So an exceedingly complex adjustment<br />

is made of the unit cost applicable to such services. Modification<br />

coefficients are currently classified according to eight categories.<br />

For example, if a local government has a smaller population with<br />

diseconomies of scale, adjustments would have to be made for the<br />

additional cost of public services. In conclusion, while there<br />

continues to be strong support for addressing local expenditure<br />

needs in Japan, it adds some complexity to the programme. This<br />

would call for a simplification of the formula aiming at improving<br />

transparency, awareness, and understanding of the <strong>Local</strong><br />

Allocation Tax.<br />

7.5. Identifying the issues<br />

Objective criteria<br />

For the past decade, we have listened to a wide range of opinions<br />

and ideas about the <strong>Local</strong> Allocation Tax and how it should be<br />

changed for the future. The following provides highlights of the key<br />

themes and issues identified (Mochida 2006).<br />

Once a minimum level for welfare services was achieved, the<br />

definition of the level of standard fiscal need became increasingly<br />

ambiguous. As a result, the coverage and costs of fiscal<br />

178


equalisation schemes have increased over the past decades. As a<br />

proportion of total central government spending, LAT transfers<br />

have ranged from a low of 16.8 per cent in 1984 to a high of 25.5<br />

per cent in 1999, and most local governments are now eligible for<br />

these grants. Several factors have contributed to the upward<br />

pressures on the equalisation system. First, the local allocation tax<br />

becomes asymmetric in adjusting for the business cycle (Mochida<br />

2004; OECD 2005). The funding pool—namely, a set percentage of<br />

the national tax—expands when the economy is growing. Cyclical<br />

windfall tax revenues made it possible to upgrade national<br />

standards for local public services. However, during downturns it<br />

has been difficult to cut back these transfers. The decline in the<br />

funding pool for the LAT has largely been compensated for by<br />

borrowing from the LAT special account or by encouraging local<br />

governments to issue bonds, for which the future repayment costs<br />

are partly included in the calculation of entitlement of the LAT.<br />

Second, several observers have argued that the vague definition of<br />

public service included in the standard fiscal needs has contributed<br />

to ballooning demands for transfers. Concerns have been expressed<br />

that the rapid expansion of standard fiscal needs over the past<br />

decades reflects efforts to stimulate the economy through specific<br />

incentives, rather than expenditure needs per se.<br />

The source of political influence or discretionary elements lies in<br />

the composition of the <strong>Local</strong> Public Finance Plan. It is noteworthy<br />

that not only all of the obligatory expenditures but also most of the<br />

discretionary expenditures are involved in LPFP. As to the former,<br />

local governments are by law obliged to provide them, which is why<br />

there are objective criteria for them. Teacher allocations and the<br />

number of students per class are clearly spelled out in the<br />

conditions attached to the specific purpose grants, and standards of<br />

public assistance are determined by the central government.<br />

However, there is room for manipulating the coverage and levels of<br />

’national standards’ for discretionary expenditure with no<br />

conditions attached. As can be seen from figure 2, while the major<br />

composition of the <strong>Local</strong> Public Finance Plan relates to obligatory<br />

expenditure highlighted in dark grey, discretionary expenditures<br />

such as waste treatment, social welfare, independent public works<br />

are also involved in the LPFP with a substantial volume<br />

(highlighted in light grey).<br />

179


Figure 2. <strong>Local</strong> public finance programs (FY2005): 83.76 trillion<br />

yen<br />

(Unit: 100 million yen)<br />

Subsidized projects<br />

66,463<br />

Salary-related<br />

expense<br />

Independent<br />

227,240<br />

projects<br />

160,777<br />

Subsidized projects<br />

100,538<br />

General<br />

administrative<br />

Independent<br />

expense<br />

projects<br />

122403<br />

231,307<br />

Investment<br />

expense<br />

196,761<br />

National insurance<br />

8,366<br />

Subsidized projects<br />

(Public works<br />

projects, etc.)<br />

73,061<br />

Independent<br />

projects<br />

123,700<br />

Public bond expence<br />

133,803<br />

Public enterprise<br />

28,659<br />

Other<br />

19,917<br />

National grants: 21,448<br />

<strong>Local</strong> burden: 45,015<br />

58,737<br />

Family register desk,<br />

welfare offices,<br />

helathcare centers,<br />

waste treatment,<br />

food service centers,<br />

etc.<br />

102,040<br />

National grants: 46,583<br />

<strong>Local</strong> burden: 53,955<br />

Investments to national corporations,<br />

agencies, etc.<br />

3.703<br />

3,703<br />

<strong>Social</strong> welfare system<br />

expenses<br />

Other<br />

<strong>Local</strong> burden<br />

45,173<br />

Amount borne for goverment-controlled<br />

project: 11,351<br />

National grants: 32,675<br />

<strong>Local</strong> burden: 29,035<br />

13,380<br />

Other<br />

Interest subsidies: 3<br />

<strong>Local</strong> bonds 133,800<br />

Items related to redemption<br />

of principal and interest for<br />

20,568<br />

enterprise loans<br />

Other than the above 8,091<br />

Elementary and junior high school teachers<br />

Police officers: 23,974<br />

Firefighters: 13,087<br />

High school teachers: 21,676<br />

Livehood Livelihood support, long-term care insurance<br />

(nursing homes, home helpers, etc.),<br />

medical care for the elderly<br />

(partially borne y public bonds), etc.<br />

childcare (nursery schools, tec.)<br />

etc.) etc.<br />

Operational expenses for police and firefihtting,<br />

waste treatment, management of roads and rivers,<br />

commercial and industrial loans, public health<br />

centers, school operation expense, private school<br />

subsidies.<br />

Temporary grants for maintenance of regional roads<br />

Cleaning; agriculture, forestry, and fiseries, rivers<br />

and coasts, urban planning, education,etc.<br />

waterworks Waterworks and sewarage, hospitals (advanced<br />

medical care, etc.)<br />

Source: “Publication of of the <strong>Local</strong> Public Finance Bureau, Bureau”, Ministry of of Internal Affairs and<br />

and<br />

Communications (2005b)<br />

Policy neutrality and sound incentives<br />

The fiscal equalisation system should be as neutral as possible and<br />

should not distort the autonomous fiscal management by local<br />

governments. Most experts agree that, on grounds of policy<br />

neutrality, differences in expenditure needs should be equalised<br />

only to the extent that they are beyond the control of government.<br />

180


Otherwise, equalisation distorts incentives for sound public<br />

management. On the other hand, fiscal transfer that covers gaps<br />

between actual expenditure and actual revenue is not desirable<br />

because it gives rise to moral hazards in local governments. In this<br />

regard, Japan’s equalisation system has both strengths and<br />

weaknesses. The calculation formula for the local allocation tax<br />

contains a number of attractive features that are intended to<br />

contain inefficient behaviour. Because it is calculated based on<br />

standardised tax revenues, local allocation tax entitlements are not<br />

affected, even if the actual tax rate changes. Furthermore, 25 per<br />

cent of the estimated tax revenue is not included in the fiscal<br />

equalisation system. This gives local governments an incentive to<br />

expand their tax base to some extent.<br />

However, current LAT transfer hinders local governments’<br />

incentive to provide efficient services in a number of areas. The<br />

main problem areas are the following:<br />

• The expenditures of repaying capital and interest on local<br />

bonds are included in the standard fiscal needs of the local<br />

allocation tax. It is done through the modification coefficient<br />

for debt services (that is, through the calculation of<br />

investment expenses by converting population, area, etc. as<br />

well as the amount of local bonds issued into measurement<br />

units). For example, in comprehensive regional<br />

improvement projects, 75 per cent of the financial resources<br />

are procured through local bonds, but 55 per cent of the<br />

future repayment costs are included in the LAT<br />

entitlement. Mochida (2004) estimated that 36 per cent of<br />

the total outstanding local debt amounting to 130 trillion<br />

yen will be repaid by the central government. This provided<br />

an incentive for local governments to incur higher levels of<br />

debts.<br />

• <strong>Local</strong> allocation tax gives premiums to small local<br />

governments (called modification coefficients for local<br />

government size). Originally, it is intended to take into<br />

account diseconomies of scale where fixed costs must be<br />

spread among a smaller population. However, these<br />

generous premiums are criticised by The Council of<br />

Economic and Fiscal Policy for obstructing efforts towards<br />

administrative reform and mergers. Consequently, the<br />

181


182<br />

modification coefficients for local government size were<br />

scaled down in the fiscal year 2002.<br />

• As discussed before, the prescribed percentage of the<br />

estimated tax revenue is not included in the fiscal<br />

equalisation system. The reservation rate gives local<br />

governments an incentive to expand their tax base.<br />

However, faced with incredible debt accumulations, The<br />

Council of Economic and Fiscal Policy (CEFP) severely<br />

criticised it for creating a ‘poverty trap’ and establishing a<br />

permanent sense of entitlement to LAT. Consequently, the<br />

central government reluctantly raised the reservation rate<br />

for the prefectures from 20 per cent to 25 per cent in 2002.<br />

However, other observers argue that this criticism does not<br />

have sufficient empirical evidence (Horiba, Mochida, and<br />

Fukae 2003). We argue that it is systematically impossible<br />

for local governments to get more LAT transfers by lowering<br />

the effective tax rate or by making fewer efforts to expand<br />

their local economy, because they are not allowed a wide<br />

range of discretion in setting rates or changing tax bases.<br />

Affordability and stability<br />

The funding pool for fiscal equalisation should be stable so that<br />

local governments can prepare appropriate budgets. However, the<br />

funding pool must be combined with some degree of flexibility to<br />

ensure that nothing prevents macroeconomic stabilisation. A way<br />

of resolving these conflicting demands is to fix the funding pool at a<br />

certain percentage of national revenue and to review this rate<br />

through regular negotiations (Litvack, Ahmad, and Bird 1998).<br />

Japan’s equalisation has both strengths and weaknesses.<br />

Approximately 32 per cent of five national taxes (income tax on<br />

individuals, corporate income tax, value added tax, liquor tax, and<br />

cigarette tax) are secured by law as a funding pool for the local<br />

allocation tax. One merit of this method is that abundant resources<br />

are provided to the LAT transfer during periods of rapid economic<br />

growth such as in 1950-early70s. However, the funding pool is<br />

sensitive to economic fluctuations and thus tends to expand or<br />

contract. Unfortunately, the income tax on individuals and<br />

corporate income tax, which form part of the funding pool for the<br />

local allocation tax, have overlapping tax bases with the inhabitant<br />

tax and enterprise tax, which are important taxes for local


governments. All of these taxes have a truly correlative<br />

relationship with economic fluctuations. This problem gave rise to<br />

shortages in financial resources during the “lost decade” in the<br />

1990s.<br />

Although rules for eliminating the gap between the funding pool<br />

and entitlements are established by law, such rules do not always<br />

function effectively. Article 6, paragraph 3-2, of the <strong>Local</strong><br />

Allocation Tax Law (Law 211 of 1950) addresses a situation in<br />

which a shortage of financial resources in local fiscal planning<br />

comes to roughly 10 per cent or more of the LAT transfers, and this<br />

condition has continued for two years and is predicted to continue<br />

for another year or longer. The law stipulates that, in such cases,<br />

the tax-sharing ratio will be raised. In actuality, the tax-sharing<br />

ratio has been raised gradually since 1954 when it was 20 per cent.<br />

In 1966, the shares of three national taxes were raised to 32 per<br />

cent in order to deal with financial resource shortages in local<br />

governments. In FY 1989, a portion of the VAT and the cigarette<br />

tax had been added to the funding pool. While the situation<br />

described in article 6 of the <strong>Local</strong> Allocation Tax Law has<br />

continued for eight consecutive years since 1996, the tax-sharing<br />

ratio has not been raised.<br />

7.6. Revitalising the equalisation system<br />

Improve the way of addressing expenditure needs, do not abandon<br />

it<br />

Fiscal equalisation is clearly a Japanese conundrum. There is no<br />

perfect solution for equalisation in Japan. Given the dynamics of<br />

Japan’s intergovernmental relations and the political equilibrium<br />

between the main political actors, perfection cannot be expected.<br />

Nonetheless, we should have to pursue the goal of developing a<br />

sound and effective equalisation system for Japan. What follows is<br />

a package of reforms designed to revitalise Japan’s fiscal<br />

equalisation programme and to establish a sustainable basis for<br />

the future.<br />

While there is a continued strong support for the purpose and<br />

principles underlying the LAT transfers, some commentators and<br />

analysts have called the programme into question. The Council of<br />

Fiscal System set up by the MOF argues that current LAT should<br />

183


e completely replaced by a new equalisation system that should<br />

focus only on revenue equalisation and take differences in<br />

expenditure needs into account by using specific grants. In their<br />

view, the central government budget deficit is higher than that of<br />

local governments; this asymmetry requires cutting back fiscal<br />

transfers including the LAT. They also argue that the national<br />

minimum has already been reached in most public services; local<br />

governments should tailor their outputs of goods and services to<br />

the particular preferences of their constituencies. In their view, the<br />

LAT obstructs efforts towards cutting wasteful spending, because<br />

that would fill the full gap between expenditures and revenues<br />

regardless of the differences in the efficiency and effectiveness of<br />

government service delivery.<br />

We do not believe that a complete replacement is appropriate for a<br />

number of reasons:<br />

First, the primary balance of local governments, by international<br />

comparison, is relatively high, and their outstanding debt has<br />

soared rapidly over the past decade. If the government were to cut<br />

back fiscal transfers substantially, this would simply replace the<br />

central government’s fiscal deficit with local governments’ debts<br />

(the Council of <strong>Local</strong> Finance 2006).<br />

Second, if sub-national governments were to have considerable<br />

autonomy and responsibility for the vast majority of public services<br />

like in Canada, the LAT’s function of guaranteeing adequate<br />

revenues to provide a uniform level of public services would be<br />

almost unnecessary. To the extent that the differences in tax rates<br />

correspond to differences in local service levels, the tax<br />

differentials are in conformity with the principle of benefit<br />

taxation. However, the LAT’s function of guaranteeing adequate<br />

revenues needs to be retained in the vast majority of public<br />

services including compulsory education, health, and social<br />

services in Japan, where equal access to public service and the fair<br />

burden of it can be viewed as essential elements of the<br />

development.<br />

Third, the LAT does not fill the gap between actual expenditures<br />

and actual revenues as some critics comment; instead, the<br />

entitlement to LAT transfers is basically determined by objective<br />

184


variables that lie beyond the control of local governments. In<br />

conclusion, the addressing expenditure needs in equalisation<br />

should be retained; doing so will require improvements in the way<br />

expenditure needs are addressed. In this way, all local<br />

governments continue to be strong supporters of the LAT, as are<br />

many academics and experts.<br />

Formula-driven transfers with simplicity and transparency<br />

The longstanding upward drift in the coverage and costs of the<br />

LAT needs to be halted. The cut in the LAT of 5 trillion yen over<br />

the period FY 2001-2006 is a step in the right direction.<br />

Recognising that the quality of public services is now high in most<br />

parts of the country, there is a need to reassess the coverage and<br />

generosity of the LAT against other major national policy<br />

objectives, in particular the need to restore fiscal sustainability.<br />

The coverage of “standard fiscal need”, which enters into the<br />

calculation of grant entitlements, should be systematically<br />

reviewed and be open to public debate. Some commentators and<br />

analysts have estimated the ratio of discretional expenditure over<br />

the total amount of the <strong>Local</strong> Public Finance Plan, with results<br />

ranging from 10 per cent to 40 per cent. In particular, the inclusion<br />

of some public infrastructure endowments, of which the impacts on<br />

local resident welfare seem low, should be reconsidered. In<br />

parallel, the funding pool available for equalisation should be<br />

raised as stipulated by Article 6, paragraph 3-2, of the <strong>Local</strong><br />

Allocation Tax Law (Law 211 of 1950). This would help avoid<br />

discretionary elements in determining the size of the funding pool.<br />

The LAT formula should be redesigned so as to encourage local<br />

governments to deliver services in the most cost-effective manner.<br />

First, it should be based more systematically on objective criteria<br />

(e.g. size of school age population, remoteness indices) rather than<br />

actual endowments, such as public infrastructure. This would<br />

mitigate the incentives for local governments to embark on<br />

spending with low economic returns or welfare impacts.<br />

The reliance on discretionary elements, including special LAT and<br />

various aspects of modification coefficients used in the LAT<br />

calculation, should be reduced, and, in general, the formula should<br />

be made more transparent. Since FY2007, 10 per cent of the total<br />

185


standard fiscal needs were to be assigned to each local government<br />

on the basis of the number of population and the size of its area.<br />

The portion of the so-called “new-model <strong>Local</strong> Allocation Tax” was<br />

to increase to 30 per cent in the near future. This reform presents<br />

advantages to local government, because it would contribute to a<br />

simplification of the formula. But the results of this ‘new formula’<br />

would not always be precise, because they reflect only a relative<br />

demand for public services, or “workload”, which is potentially<br />

different from the real picture. Put differently, the new formula<br />

neglects the financial costs of providing a given set of services. For<br />

example, when population is used as an indicator, remote areas<br />

with relatively high costs would not be able to provide adequate<br />

public services.<br />

Similarly, there is a need to remove existing incentives in the LAT<br />

system in favour of local bond issues, because they may hinder<br />

otherwise desirable adjustments in local tax rates or cuts in<br />

spending. The reduction in the LAT coverage of repayment costs<br />

for some public infrastructures made in FY 2002 is a step in the<br />

right direction, as is the increase in the share of local taxes<br />

retained by prefectures. A bolder reform, however, is needed since<br />

a large share (commonly between 35 and 100 per cent) of local<br />

bond repayment costs is still repaid by the equalisation scheme,<br />

while 75 per cent of a rise in local tax revenues stemming from an<br />

increase in taxable income is given away out of the equalisation<br />

pool.<br />

References<br />

Akizuki, Kengo. 2001. “Partnership in Controlled Decentralization: <strong>Local</strong><br />

<strong>Government</strong>s and the Ministry of Home Affairs.” In <strong>Local</strong> <strong>Government</strong><br />

Development in Post-War Japan, ed. Michio Muramastu, Farrukh Iqbal,<br />

and Ikuo Kume, Ch.3,63-84. Oxford University Press.<br />

Boadway, Robin W., Paul A. Hobson, and Nobuki Mochida. 2001. “Fiscal<br />

Equalization in Japan: Assessment and Recommendations.” Journal of<br />

Economics (University of Tokyo) 66 (4): 24–57.<br />

Buchanan, James M. 1950. “Federalism and Fiscal Equity,” American Economic<br />

Review Vol. 40, 583-99.<br />

General Headquarters Supreme Commander for Allied Power. 1949. Report on<br />

Japanese Taxation. GHQ, Tokyo.<br />

Horiba, Isao, Nobuki Mochida, and Keishi Fukae. 2003. “Moral Hazard Effect of<br />

<strong>Local</strong> Allocation Tax on Property Tax” (in Japanese). Aoyamagakuin<br />

Journal of Economics 54 (4): 27-58.<br />

186


Kitamura, Wataru. 2006. “Trinity Reform and its Impact on Intergovernmental<br />

Relations” (in Japanese). in Institute of <strong>Social</strong> Science eds, Beyond the ‘Lost<br />

Decade’: the Era of Koizumi’s Reform, Tokyo, University of Tokyo Press,<br />

219-246.<br />

Litvack Jennie, Junaid Ahmad, and Richard Bird. 1998. Rethinking<br />

Decentralization in Developing Countries. Washington, DC: The World<br />

Bank.<br />

Lotz, Jørgen. 2006. “Grants and Equalization; a Nordic perspective,” paper<br />

presented for the workshop on Fiscal Equalization in the Drive to<br />

Decentralize: Global Trend in Sweeping Reform,’ held in Tokyo, Japan,<br />

November 12-13, 2005.<br />

Ma, Jun. 1996. “Intergovernmental Fiscal Transfer: A Comparison of Nine<br />

Countries(Cases of the United States, Canada, the United Kingdom,<br />

Australia, Germany, Japan, Korea, India, and Indonesia),” paper prepared<br />

for the workshop on ‘Intergovernmental Fiscal Transfer: The Experiences<br />

from Japan and other Countries,’ held in Hanoi, Vietnam, March 27-<br />

29,1996.<br />

MIC (Ministry of Internal Affairs and Communication). 2005b.Materials on <strong>Local</strong><br />

Tax System,(in Japanese), MIC, Tokyo, unpublished.<br />

Mochida, Nobuki. 1998. “An Equalization Transfer Scheme in Japan,” in Shibata/<br />

Ihori (eds.), Welfare State, Public Investment and Growth, Springer-Verlag<br />

Tokyo,269-293.<br />

———. 2001. “Taxes and Transfers in Japan’s <strong>Local</strong> Finances.” In <strong>Local</strong><br />

<strong>Government</strong> Development in Post-War Japan, ed. Michio Muramastu,<br />

Farrukh Iqbal, and Ikuo Kume, Ch4,85-111. Oxford, U.K.: Oxford<br />

University Press.<br />

———. 2004. Fiscal Decentralization and State-<strong>Local</strong> Finance: Fundamental<br />

Perspectives (in Japanese). Tokyo: University of Tokyo Press.<br />

———. 2006. “<strong>Local</strong> <strong>Government</strong> Organization and Finance in Japan,” in Anwar<br />

Shah ed, <strong>Local</strong> Governance in Industrial Countries, Washington: World<br />

Bank, 149-188.<br />

———. 2008. Fiscal Decentralization and <strong>Local</strong> Public Finance in Japan,<br />

Routledge, forthcoming.<br />

Mochida, Nobuki, and Jørgen Lotz. 1999. “Fiscal Federalism in Practice: the<br />

Nordic Countries and Japan.” Journal of Economics (University of Tokyo)<br />

64 (4): 55–86.<br />

Muramatsu, Michio, and Farrukh Iqbal. 2001. “Understanding Japanese<br />

Intergovernmental Relations: Perspectives, Models, and Salient<br />

Characteristics.” In <strong>Local</strong> <strong>Government</strong> Development in Post-War Japan, ed.<br />

Michio Muramastu, Farrukh Iqbal, and Ikuo Kume, Ch1,1-28. Oxford, U.K.:<br />

Oxford University Press.<br />

OECD (Organisation for Economic Co-operation and Development). 1999. “Taxing<br />

Powers of State and <strong>Local</strong> <strong>Government</strong>.” OECD Tax Policy Studies 1,<br />

OECD, Paris.<br />

———. 2005. Economic Surveys: Japan. Paris: OECD.<br />

Policy Research Institute, Ministry of Finance. 2002. Comparative Study of <strong>Local</strong><br />

Finance System (in Japanese). Tokyo: Policy Research Institute.<br />

Shah, Anwar. 1994. “The Reform of Intergovernmental Fiscal Relations in<br />

Developing and Emerging Market Economies.” Policy and Research Series<br />

23, World Bank, Washington DC.<br />

187


Chapter 8<br />

Annual fluctuations in the cost equalisation<br />

and budget stability for the municipalities in<br />

Sweden<br />

Lennart Tingvall<br />

Summary<br />

The cost equalisation for the Swedish municipalities, used since<br />

1996, is built up of some 10 distinct models. The different models<br />

refer to services like child care, elderly care and compulsory school<br />

etc. Many of the factors on which the equalisation is based are<br />

updated annually. The age structure and the factors that reflect<br />

the municipalities’ socio-economic structure are thus updated<br />

annually. It is perfectly natural that this updating leads to a<br />

variation in the result of cost equalisation for the individual<br />

municipality from one year to the next. The annual variations can<br />

be large for municipalities with a low population, which makes<br />

budgeting difficult. There is a conflict between stability and<br />

changes in the cost equalisation that reflect changes in<br />

expenditure needs. However, the largest changes have happened<br />

when the system has been followed up, which has taken place<br />

every 4-5 years. This paper concentrates on the equalisation for<br />

municipalities, but the system for the county councils is partly<br />

described in brief.<br />

8.1. Why is local government equalisation necessary<br />

Public expenditure in Sweden accounts for a high share of GDP,<br />

and the scale of decentralisation of tasks to the 290 municipalities<br />

and the 20 county councils is substantial in international terms.<br />

Because municipalities and county councils account for the bulk of<br />

our welfare services by being responsible for schools, health care<br />

and other forms of care, local government expenditure accounts for<br />

a higher share of GDP than in any other country. At the same<br />

188


time, conditions vary greatly between local authorities and<br />

different parts of the country, as do conditions for fulfilling these<br />

tasks.<br />

On average, some 70 per cent of municipal and county council costs<br />

for services are financed by local income tax. Only a small<br />

proportion of the revenues, 15 per cent on average including the<br />

equalisation grants, consists of central government grants. But<br />

local authorities differ in their ability to meet their tasks because<br />

there are major differences in taxable income per inhabitant (tax<br />

capacity) between municipalities and between county councils.<br />

This is due not only to differences in inhabitants’ income from<br />

work but also to differences in employment levels and in the age<br />

structure of their population.<br />

In 2005, the municipality in Sweden with the highest tax capacity<br />

had 174 per cent of the national average tax capacity<br />

(corresponding to an average taxable income of almost SEK<br />

267 000 for inhabitants). The municipality with lowest tax capacity<br />

had just under 79 per cent of the average tax capacity (SEK<br />

121 000 per inhabitant). Without an equalisation of tax capacity,<br />

the inhabitants of a low-income municipality would need to pay 26<br />

per cent of their income in municipal tax, while people living in a<br />

high income municipality would only need to pay 12 per cent in<br />

municipal tax to give the two municipalities the same tax revenue<br />

per inhabitant as the national average.<br />

This is a calculation example, but the difference in municipal and<br />

county council taxes between residents of a low-income<br />

municipality and a high-income municipality would be almost 17<br />

per cent of their income or close to SEK 26 000 for a middle-income<br />

earner in the two municipalities.<br />

In addition, on account of differences in the need for local government<br />

services, partly due to differences in population age structure, the cost<br />

differences between local authorities of delivering an equivalent level<br />

of services to inhabitants may be considerable.<br />

189


Figure 1. Average municipality cost per year for persons of<br />

different ages<br />

300000<br />

Other costs Childcare and education Care of the elderly<br />

250000<br />

SEK per inhabitant<br />

200000<br />

150000<br />

100000<br />

50000<br />

0<br />

0<br />

4<br />

8<br />

12<br />

16<br />

20<br />

24<br />

28<br />

32<br />

36<br />

40<br />

44<br />

48<br />

52<br />

56<br />

60<br />

64<br />

68<br />

72<br />

76<br />

80<br />

84<br />

88<br />

92<br />

96<br />

Age<br />

100<br />

The costs of elderly care in the municipality that has the lowest<br />

proportion of elderly in the country has been calculated in cost<br />

equalisation as some SEK 3 000 per municipal inhabitant. In the<br />

municipalities that have the highest proportion of elderly these<br />

costs have been calculated as almost SEK 17 000 per municipal<br />

inhabitant. This difference corresponds to 9 per cent of the<br />

municipal tax, based on the national average tax base per<br />

inhabitant. Without equalisation, a rural municipality with lots of<br />

elderly people in need of care would have difficulty in providing<br />

elderly care unless it levied a very high municipal tax.<br />

In Sweden there is a long-standing, broad political agreement that<br />

people should have equal access to welfare, no matter where in<br />

Sweden they live. These two examples point to the need for<br />

financial equalisation, both for tax revenue and for structural cost<br />

differences, in order to put municipalities and county councils on<br />

an equal financial footing for the conduct of their activities. The<br />

redistribution of resources that takes place in the equalisation<br />

system makes it possible to establish more equal conditions for the<br />

provision of welfare services.<br />

190


Different systems for equalisation over the years<br />

For most of the 20th century, municipalities and county councils<br />

have received some form of state grant in order to equalise for the<br />

differences in their financial situations. However, an equalisation<br />

system in the proper sense of term was introduced in 1966. The<br />

system consists of two parts: income equalisation along with a<br />

separate equalisation grant to municipalities and county councils<br />

with high tax rates, or a reduced grant for municipalities and<br />

county councils with low tax rates. There were a number of specific<br />

grants in place alongside the equalisation system. The main<br />

features of this system were to remain in force until the end of<br />

1992, even though the system was modified on a number of<br />

occasions.<br />

The 1993 reform of local government finances<br />

In 1993, an extensive reform was made of the systems of central<br />

government grants. Several specific government grants were<br />

replaced by one general grant, i.e. “all the money in one pot.” The<br />

new equalisation system for municipalities consisted of three parts:<br />

income equalisation, cost equalisation and a supplement for<br />

municipalities with a large population decrease. However, criticism<br />

was directed at the cost equalisation component in particular, and<br />

a new inquiry was therefore set up to review it. An additional<br />

problem was that municipalities with a very high tax capacity level<br />

fell outside the grant system. Thus, the government grant could<br />

not be used for general financial adjustments between the State<br />

and municipalities.<br />

1996 – a new government grant and equalisation system<br />

The 1996 government grant and equalisation system consisted of<br />

four parts: income equalisation, cost equalisation, a general<br />

government grant and transitional regulations. Unlike what hasd<br />

been the case of previous systems, all municipalities and county<br />

councils were now included. Equalisation was now also neutral in<br />

terms of central government finances, as both income and cost<br />

equalisation were financed by municipalities and county councils –<br />

both equalisations now became horizontal. Municipalities and<br />

county councils with a tax capacity below and/or structural costs<br />

above the average level received grants while other municipalities<br />

and county councils had to pay a charge. In principle, the national<br />

total of charges was equal to the total amount of the grants.<br />

191


Cost equalisation was designed according to the same principles in<br />

both the municipal and county council sectors. The method is<br />

called the standard cost method. It means that cost equalisation is<br />

built up of a number of models calculated with the use of various<br />

factors (table 3). The various models refer either to specific services<br />

or to costs found in most services.<br />

2000 – Reviews of the 1996 system<br />

From 2000, certain changes were made in the cost equalisation<br />

component of the system.<br />

2005 – New changes<br />

A new system of local government financial equalisation was<br />

introduced in Sweden on 1 January 2005. It has the same purpose<br />

as the previous system: to put all municipalities and county<br />

councils in the country on an equal financial footing to deliver<br />

equivalent levels of services to their residents, irrespective of the<br />

income of local authority residents and other structural factors.<br />

The intention is that differences in local tax rates should basically<br />

reflect differences in efficiency and in levels of services and charges<br />

and not be due to differences in structural conditions. Since 2005,<br />

the equalisation system has consisted of five separate parts,<br />

income equalisation, cost equalisation, structural grant,<br />

transitional grant and adjustment grant/charge.<br />

Table 1 shows the total of the grants and charges included in the<br />

new system of local government financial equalisation and the<br />

total volume of the system in 2007.<br />

Table 1. <strong>Local</strong> government tax equalisation<br />

(Calculations for 2007, SEK billion)<br />

Municipalities County councils Total<br />

Income equalisation grant 48.1 15.8 63.9<br />

Income equalisation charge -3.3 -2.0 -5.2<br />

Cost equalisation grant 4.9 1.2 6.0<br />

Cost equalisation charge -4.9 -1.2 -6.0<br />

Structural grant 1.5 0.7 2.2<br />

Transitional grant 0.5 0.2 0.7<br />

Total 46.8 14.7 58.7<br />

Adjustment grant 8.3 0.9 9.3<br />

Total 55.2 15.7 70.8<br />

192


The Swedish Tax Agency, which collects local tax revenue,<br />

administers the various grants and charges when it pays local<br />

government tax revenue. The Tax Agency adjusts the tax revenue<br />

for any grants or charges.<br />

2008 – New changes again<br />

From 2008, a number of new changes to the cost equalisation will<br />

be introduced. A new component from 2008 is structural<br />

differences in wage levels between municipalities and between<br />

county councils.<br />

The central government has proposed that a new parliamentary<br />

committee is to start reviewing all parts of the equalisation<br />

system, beginning in January 2008.<br />

8.2. New system for local government financial equalisation<br />

2005<br />

Income equalisation<br />

Income equalisation results in an equalisation of tax revenue<br />

between municipalities and between county councils. One<br />

fundamental difference compared with the previous system is that<br />

horizontal income equalisation between local authorities has been<br />

replaced by a mainly state-financed (vertical) income equalisation<br />

grant. The general central government per capita grant to<br />

municipalities and county councils was terminated at the same<br />

time.<br />

The income equalisation grant is calculated on the basis of the<br />

difference between the local authority’s own taxable income and a<br />

tax equalisation base that corresponds to 115 per cent of the<br />

national average tax capacity per capita for municipalities and 110<br />

per cent for county councils. Municipalities and county councils<br />

that have a taxable income that exceeds these levels have to pay<br />

an income equalisation charge to the central government instead.<br />

193


Figure 2. Income equalisation 2005, grant level and municipalities<br />

tax capacity per capita. Average tax capacity = 100 %<br />

180%<br />

115%<br />

80%<br />

Grant 95% 275 municipalities<br />

Fee 85%,<br />

15 munic.<br />

0%<br />

For grant-entitled municipalities, the grant is calculated on the<br />

basis of 95 per cent of the national average tax rate in 2003, and<br />

for grant-entitled county councils that calculation is based on 90<br />

per cent of the national average tax rate in the same year. For<br />

municipalities and county councils that are liable to pay a charge,<br />

such charge is calculated on the basis of 85 per cent of the average<br />

tax rate instead. When tax rates (the county-level tax rate) are<br />

fixed in the calculation of grants and charges, account must also be<br />

taken of differences in tax shifts between municipalities and<br />

county councils in each county as a result of changes of<br />

responsibilities from and including 1991.<br />

194


Figure 3. Income equalisation 2005. SEK per inhabitant<br />

(kommuner=municipalities, landsting=county councils)<br />

Kommuner<br />

Landsting<br />

7914 till 11361 (93)<br />

6241 till 7914 (92)<br />

2338 till 3775 (7)<br />

2212 till 2338 (6)<br />

0 till 6241 (92)<br />

–15276 till 0 (13)<br />

0 till 2212 (7)<br />

–1149 till 0 (1)<br />

Stor-Göteborg<br />

Stor-Stockholm<br />

Stor-Malmö<br />

What local authorities get grants and what local authorities pay<br />

The main charge-payers are suburban municipalities in the<br />

Stockholm County. The Stockholm Municipality is the only one of<br />

the three metropolitan cities to pay a charge. Rural municipalities<br />

and other small municipalities receive the largest grants per<br />

inhabitant. In the income equalisation system for county councils,<br />

only the Stockholm County Council pays a charge in 2005. County<br />

councils in the rest of the county receive a grant.<br />

Cost equalisation<br />

Cost equalisation evens out structural cost differences, which can<br />

be of two different kinds. One is that the need for local authority<br />

services varies: for example, municipalities with a high proportion<br />

195


of old residents have a greater need of elderly care. The other is<br />

that the cost of producing a particular service varies: for example,<br />

schools can cost more in rural municipalities as teaching there may<br />

require smaller classes, and the pupils more often need school<br />

transport.<br />

Municipalities and county councils with an unfavourable cost<br />

structure receive a grant from the State, as they did in the<br />

previous system. Those that have a favourable structure pay a<br />

charge to the State. Cost equalisation is still neutral in terms of<br />

central government finances, as the total grants and charges are<br />

equal in size and therefore cancel one another out.<br />

Cost equalisation equalises for costs that the local authority itself<br />

cannot influence. Measurable and objective factors are used to<br />

calculate a structural measure of the differences in conditions and<br />

needs.<br />

One basic principle is only to equalise for structural differences<br />

that can arise in services that are mandatory for municipalities<br />

and county councils. In addition, cost equalisation is only to apply<br />

to cost and need differences in local authority activities and not to<br />

differences in inhabitants’ private consumption, such as high<br />

housing costs.<br />

Cost equalisation is designed according to the same principles in<br />

both the municipal and the county council sector. The method that<br />

was introduced in 1996 is called the standard cost method. It<br />

means that cost equalisation is built up of a number of distinct<br />

models. The different models refer to services like childcare or<br />

elderly care.<br />

In the equalisation system that was introduced in 1993, the<br />

method that was chosen was based on analyses of the correlation<br />

between municipal expenditures and various external factors that<br />

may conceivably underlie differences in municipal costs per<br />

inhabitant. A system where the structural cost equalisation<br />

payments were based on four factors was approved.<br />

• climate<br />

• population density<br />

• age structure<br />

196


• social structure<br />

Yardsticks for these variables were established and weighed<br />

together into a uniform structural index.<br />

In practice it proved difficult to find indicators that were<br />

theoretically viable, intuitively understandable and at the same<br />

time practically applicable. This design of the system was very<br />

sharply criticised, and a special group of experts was therefore<br />

appointed to review this part of the system.<br />

The service-based structure of cost equalisation that was<br />

introduced in 1996 has both advantages and disadvantages. By<br />

treating each area separately it is possible to capture the very<br />

different conditions that apply to different services. For example<br />

the population density affects the expenditure needs in different<br />

ways among the local government services. The conditions for local<br />

government services are therefore very different for municipalities<br />

in different parts of Sweden. The average population density for<br />

the 33 municipalities in Skåne in southern Sweden is more than<br />

100 inhabitants/square kilometre and less than 0,5<br />

inhabitants/square kilometre in some municipalities in Lappland<br />

in northern Sweden. If all these structural differences between<br />

municipalities should be taken into consideration in only one<br />

national system, this system is one possible and likely acceptable<br />

solution. The ambition has been to identify indicators that are<br />

theoretically viable, intuitively understandable, practically<br />

applicable and generally accepted for each service within the cost<br />

equalisation.<br />

With this solution it is also possible to adjust cost equalisation in<br />

part when the conditions for a certain service change, so that other<br />

parts of the cost equalisation are not affected. For the<br />

municipalities, new responsibilities can be added to the system,<br />

and it is possible to move responsibilities from the county councils<br />

to the municipalities. It is also possible within the system to<br />

compensate for certain structural differences that cannot be<br />

attributed directly to special programmes, such as those related to<br />

climate or high-unit costs due to a rapid decline in population.<br />

197


The main disadvantage is that this equalisation is very<br />

comprehensive with a large number of factors and separate<br />

models. On the other hand, a simpler system risks missing certain<br />

structural costs, which would result in a less fair system. So the<br />

design of cost equalisation is a balance between simplicity and<br />

fairness. For municipalities, nine services are included in the<br />

equalisation, while for county councils one service – health and<br />

medical care – is covered. There is also one service with shared<br />

responsibility, namely public transport.<br />

Differences in standard costs are intended to reflect structural cost<br />

differences, i.e. costs that municipalities and county councils<br />

themselves cannot control. The different services and variables in<br />

each model are set out in table 3.<br />

What does cost equalisation cover<br />

As mentioned above, cost equalisation equalises for differences due<br />

to age structure, ethnicity, socio-economic conditions and<br />

geography. The model for each service uses factors that explain the<br />

extra structural costs associated with that particular service. Table<br />

2 shows what categories of extra costs the system provides<br />

compensation for in each model.<br />

Table 2. Structural extra costs by category<br />

Service Age Ethnicity Socioeconomic Geography<br />

conditions<br />

Municipalities<br />

Pre-school services and outof-school<br />

care<br />

• • •<br />

Compulsory school and preschool<br />

classes<br />

• • •<br />

Upper secondary school • • •<br />

Care of the elderly • • • •<br />

Individual and family care (•) • • •<br />

Children with a foreign<br />

background<br />

(•) •<br />

Population change<br />

(•)<br />

Settlement structure<br />

•<br />

County councils<br />

Health and medical care • • •<br />

Joint service<br />

Public transport<br />

•<br />

Note: (•) means that age is included as a factor but that it is not the primary<br />

objective of equalisation for that particular model.<br />

198


Services included in cost equalisation<br />

The standard cost in the different models is obtained by<br />

multiplying the various variables by the national average cost for<br />

each service.<br />

Table 3. Models in cost equalisation for municipalities and county<br />

councils<br />

Service<br />

Structural factors<br />

Municipalities<br />

Pre-school services and outof-school<br />

care<br />

and population density.<br />

Age structure, parents’ activity rate, tax capacity<br />

Compulsory school and preschool<br />

classes<br />

rural area.<br />

Age structure, children with a foreign background,<br />

Upper secondary school Age structure, programme choice, settlement<br />

structure<br />

Care of the elderly<br />

Age structure, sex distribution, professional<br />

background, civil status, non-Nordic background<br />

and rural area.<br />

Individual and family care • Refugees born abroad and close relatives, other<br />

people born abroad from countries outside the<br />

Nordic region and the EU, unemployed people<br />

without benefit, single women with children,<br />

proportion of men with low incomes and<br />

settlement density.<br />

• Children of lone parents, young people<br />

prosecuted, children with a foreign background<br />

and local authority population.<br />

Children with a foreign Children aged 0-19 years with a foreign<br />

background<br />

background.<br />

Population change<br />

• Population reduction > 2% in the past 10 years.<br />

• Change (positive and negative) in the number of<br />

school pupils.<br />

• Compensation for revenue delay in event of<br />

population increase.<br />

Settlement structure • Heating<br />

• Streets and roads<br />

• Building costs<br />

• Rural-specific extra costs for administration,<br />

travel and rescue services<br />

County councils<br />

Health and medical care Care-demanding groups, sex, age, civil status,<br />

employment status, income and type of housing.<br />

Supplement for rural areas.<br />

Joint service<br />

Public transport<br />

Sparseness, work commuting and urban structure.<br />

199


The standard cost for compulsory schooling is obtained by<br />

multiplying the proportion of children of school age by the national<br />

average cost per pupil and by multiplying the proportion of pupils<br />

born outside Sweden, Norway and Denmark by the average cost for<br />

mother tongue instruction. Moreover, additional costs are<br />

calculated for small schools and school transports.<br />

Cost equalisation does not and is not intended to take account of<br />

the actual costs held by municipalities or county councils. The<br />

standard cost is the cost that the municipality or county council<br />

would have if it ran the service at an average cost level, taking<br />

account of its own structural factors under cost equalisation.<br />

The use of the national average cost means that equalisation is<br />

carried out to the average level of service, quality, charge and<br />

efficiency. If a municipality or county council has higher costs<br />

because it provides a higher than average service level, this must<br />

be financed though higher taxes, more efficient services or a higher<br />

level of charge-financing. In the case of lower costs, the opposite is<br />

true. No local authority is “punished” or “favoured” because it has<br />

a cost level that differs from the average cost level.<br />

Table 4 shows that more than 85 per cent of the standard costs<br />

refer to services that are very dependent on the age structure in<br />

the municipality.<br />

Table 4. Standard costs in cost equalisation SEK per inhabitant<br />

municipalities in the models used in cost equalisation 2007<br />

Model<br />

Standard cost in SEK per inhabitant<br />

Average Per cent of total<br />

Child care 4903 17.1<br />

Compulsory school 8640 30.1<br />

Upper secondary school 3345 11.7<br />

Individual and family care 2927 10.2<br />

Children with a foreign background 80 0.3<br />

Care of the elderly 7958 27.7<br />

Population change 158 0.5<br />

Settlement structure 173 0.6<br />

Public transport 526 1.8<br />

Total 28713 100<br />

200


The equalisation system for municipalities has for several decades<br />

contained a compensation for population reduction. The purpose<br />

has been to compensate the municipalities for fixed capital costs<br />

that cannot be reduced in the same rate as the reduction of<br />

inhabitants and revenues. Another reason for the compensation is<br />

the pension liabilities that will be more expensive per capita with a<br />

reduced population in the future. The municipalities are<br />

compensated when the population reduction in the past 10 years<br />

has been above 2 %. Municipalities are also compensated for rapid<br />

changes (positive and negative) in the number of school pupils.<br />

Municipalities with a rapid population increase have since 2005<br />

(and county councils since 2006) been compensated within the<br />

equalisation system for revenue delay.<br />

Some parts of the cost equalisation are partly based on the number<br />

of inhabitants (Rural-specific extra costs for administration, travel<br />

and rescue services). Other parts of the cost equalisation are based<br />

on population density (inhabitants/square kilometre or other types<br />

of settlement densities). In the calculation of standard costs for the<br />

following services, rural areas or other types of settlement<br />

densities are included: Pre-school services, Compulsory school,<br />

Upper secondary school, Individual and family care, Health and<br />

medical care (county councils) and Public transport (municipalities<br />

and county councils). A special structural grant is partly<br />

calculated based on the number of inhabitants and population<br />

density.<br />

8.3. Calculation of grant or charge in cost equalisation<br />

The total of the standard costs for all the models is called the<br />

municipality’s or county council’s structural cost. If the structural<br />

cost is above the national average, the municipality or county<br />

council receives a grant corresponding to the difference compared<br />

to the average. The municipality or county council is regarded as<br />

having an unfavourable structure in relation to other<br />

municipalities or county councils respectively. Table 5 shows the<br />

highest and lowest structural costs in 2007.<br />

Table 6 shows the differences in standard costs among the<br />

different models. The largest difference occur in the model for care<br />

of the elderly, almost 14 000 SEK per municipal inhabitant. For<br />

201


compulsory school, child care and individual and family care the<br />

differences are around 6 600, 5 700 and 4 100 SEK respectively.<br />

Table 5. The highest and lowest municipal structural cost and the<br />

grant/charge in 2007. SEK per municipal inhabitant<br />

Municipality<br />

Grant/charge<br />

Max. 38 835 10 122<br />

Min. 24 872 –3 841<br />

Average 28 713 0<br />

Table 6. Differences in standard costs for municipalities in the<br />

models used in cost equalisation. SEK per inhabitant 2007<br />

Model<br />

Standard cost in cost equalisation SEK per<br />

inhabitant<br />

Average Minimum Maximum Difference<br />

Child care 4903 2713 8403 5690<br />

Compulsory school 8640 5070 11676 6606<br />

Upper secondary school 3345 1746 5135 3387<br />

Individual and family care 2927 1046 5133 4087<br />

Children with a foreign 80 0 799 799<br />

background<br />

Care of the elderly 7958 3018 16872 13844<br />

Population change 158 0 1666 1666<br />

Settlement structure 173 -140 1966 2106<br />

Public transport 526 0 1035 1035<br />

Total 28713 24872 38835 13963<br />

If the structural cost is below the national average, the<br />

municipality or county council is regarded as having a favourable<br />

structure in relation to other municipalities or county councils<br />

respectively and therefore pays a corresponding charge.<br />

How much is redistributed<br />

The size of the redistribution in the various models depends both<br />

on the cost of the service concerned and the scale of the structural<br />

differences. Elderly care is the municipal service that is most<br />

costly, and most money is also redistributed in that model.<br />

Individual and family care costs less than half as much as elderly<br />

care. However, in this area there are major structural differences<br />

that push up the amount redistributed in the model. The sum total<br />

of the redistribution made in the various models exceeds the total<br />

redistribution in cost equalisation. This is because a municipality<br />

can receive a supplement in one model at the same time as it gets a<br />

deduction in another model. No municipality receives supplements<br />

202


in every model. A few municipalities get a deduction in every<br />

model. Each model is self-financed – that is to say that the<br />

supplements and deductions cancel one another out.<br />

Table 7. Redistribution and number of grant and charge<br />

municipalities in the models used in cost equalisation<br />

Model SEK bn Number of municipalities with<br />

supplement Deduction<br />

Children with a foreign background 0.45 25 265<br />

Public transport 0.90 58 232<br />

Population change 0.96 153 137<br />

Settlement structure 1.31 86 204<br />

Upper secondary school 1.62 239 51<br />

Child care 3.37 57 233<br />

Compulsory school 3.76 215 75<br />

Individual and family care 4.01 31 257<br />

Care of the elderly 6.16 205 85<br />

Net total 4.70 159 131<br />

What authorities pay and what authorities receive grants<br />

In simple terms, redistribution takes place in cost equalisation<br />

from southern Sweden (excluding the Stockholm region) to the<br />

north of the country. In northern Sweden, the main recipients of<br />

major grants are rural municipalities in the interior part of the<br />

country and other small municipalities, while municipalities along<br />

the coast receive much less grant funds or pay a charge towards<br />

cost equalisation. In southern Sweden it is mostly major cities and<br />

suburban municipalities that pay a charge. There are<br />

municipalities in some parts of southeastern Götaland that receive<br />

substantial grants. In the Stockholm region, the Stockholm<br />

Municipality and certain suburban municipalities with social<br />

problems are grant recipients.<br />

Structural grant<br />

The structural grant consists of those parts of the previous<br />

equalisation system that have to do with regional policy and which<br />

are no longer to be included in cost equalisation. The structural<br />

grant has therefore been removed from cost equalisation and is<br />

intended to reinforce municipalities and county councils with a<br />

small population and/or labour market problems. This means that<br />

the structural grant is not paid to all municipalities and county<br />

councils but only to those local authorities that previously received<br />

203


supplements under certain models used in the calculation of cost<br />

equalisation and/or have experienced major revenue decreases as a<br />

result of the system change.<br />

Figure 4. Cost equalisation 2005. SEK per inhabitant . SEK per<br />

inhabitant (kommuner=municipalities, landsting=county councils)<br />

Kommuner<br />

1199 till 10051 (78)<br />

0till 1199 (77)<br />

–593 till 0 (68)<br />

–3425 till –593 (67)<br />

Landsting<br />

266 till 1298 (7)<br />

0 till 266 (6)<br />

–325 till 0 (7)<br />

–1066 till –325 (1)<br />

Stor-Göteborg<br />

Stor-Stockholm<br />

Stor-Malmö<br />

The grant is based on three factors. These are, first, the previous<br />

standard costs for business and employment promotion and for a<br />

weak population base that were included in cost equalisation up to<br />

and including 2004 and, second, compensation to municipalities<br />

and county councils that had a loss of revenue that exceeded a set<br />

level due to changes in the equalisation system from 2005.<br />

Municipalities exposed to grant reductions exceeding 0.56 per cent<br />

of their own tax base receive a structural grant corresponding to<br />

the excess. The corresponding figure for county councils is 0.28 per<br />

cent. For municipalities and county councils with an average tax<br />

204


ase per inhabitant, this grant corresponds to some SEK 850 and<br />

SEK 425 per inhabitant respectively.<br />

The structural grant is paid to 94 municipalities and 6 county<br />

councils. The grant is paid annually without any predetermined<br />

time limit and mainly goes to municipalities in the forest counties<br />

in central and northern Sweden, but also to municipalities in<br />

other parts of the country with weak employment, including<br />

Malmö. The three northernmost county councils and three county<br />

councils (including the Gotland Municipality) with small<br />

populations in Götaland also receive structural grant.<br />

Transitional grant<br />

The new system will result in major changes in revenue for many<br />

municipalities and some county councils. A special transitional<br />

grant is paid out in the period 2005-2010 to moderate the<br />

redistributive effects for those authorities that have experienced<br />

revenue decreases. This grant spreads the negative revenue<br />

change due to the system over several years.<br />

The municipalities and county councils that lose out in the new<br />

system are compensated through a transitional grant over a period<br />

of six years. This gives municipalities and county councils with<br />

negative changes a certain amount of time to adjust their costs if<br />

the reduction exceeds a particular level. The grant means that the<br />

annual income reduction must not be above 0.08 per cent of own<br />

tax base (or some SEK 120 per inhabitant) for municipalities and<br />

0.04 per cent for county councils (some SEK 60 per inhabitant).<br />

Adjustment grant and adjustment charge<br />

An adjustment grant and an adjustment charge have been<br />

introduced to ensure that the state grant is not affected by the<br />

level set in income equalisation and to ensure at the same time<br />

that the central government has control over the total cost of the<br />

equalisation system. It must also be possible to regulate changes of<br />

responsibility between the local government sector and the State.<br />

These will also be used for financial adjustments between the<br />

central government and the local government sector, for example<br />

when the local government financing principle is used<br />

to compensate municipalities and county councils for increased<br />

costs due to new duties imposed on them by the State.<br />

205


Figure 5. Structural grant 2005. SEK per inhabitant<br />

(kommuner=municipalities, landsting=county councils)<br />

Kommuner<br />

451 till 5438 (47)<br />

1 till 451 (47)<br />

0 (196)<br />

Landsting<br />

502 till 1426 (3)<br />

1 till 502 (3)<br />

0 (15)<br />

Stor-Göteborg<br />

Stor-Stockholm<br />

Stor-Malmö<br />

In the new government-financed income equalisation system, costs<br />

of the income equalisation grant will increase in line with the<br />

growth of the local government tax base. This means that the State<br />

cannot know in advance what the final total amount will be. In<br />

view of the need to be able to influence the macroeconomic scope<br />

for the local government sector and also central government<br />

finances, it should be possible to influence the total grant frame for<br />

municipalities and county councils.<br />

The way this works is as follows: if the sum of all grants minus all<br />

charges paid is less than the amount that the State has decided to<br />

transfer to municipalities or county councils, all municipalities or<br />

county councils will receive an adjustment grant corresponding to<br />

the difference. The adjustment grant is paid as a uniform amount<br />

206


per inhabitant. If, instead, the sum of all grants minus all charges<br />

paid is higher than the amount that the central government has<br />

decided to transfer, the difference has to been recovered through<br />

an adjustment charge on all municipalities and county councils,<br />

calculated in the same way as the adjustment grant. This enables<br />

the Riksdag to decide how many funds are to be allocated to<br />

municipalities and county councils.<br />

Total outcome of the equalisation system<br />

The total outcome of the equalisation system for the different<br />

municipality groups are shown in figure 6. Income equalisation<br />

charges are only paid by one (Stockholm) of the three metropolitan<br />

cities and some suburban municipalities.<br />

Rural municipalities are the largest recipients of cost equalisation<br />

grants and structural grants.<br />

Figure 6. Total outcome of the equalisation system for<br />

municipalities in 2007<br />

(SEK per inhabitant per municipality group)<br />

Income equalisation grant Income equalisation charge Cost equalisation grant<br />

Cost equalisation charge Structural + Transitional grant Adjustment grant<br />

Net grant<br />

16000<br />

14000<br />

12000<br />

SEK per inhabitant<br />

10000<br />

8000<br />

6000<br />

4000<br />

2000<br />

0<br />

-2000<br />

Metropolitan cities<br />

Suburban municipalities<br />

Large cities<br />

Commuter municip.<br />

Rural municipalities<br />

Manufacturing municip.<br />