682

Jos Roberto Afonso, Sergio Guimares Ferreira, and Ricardo Varsano,

Fiscal Competition
1. Introduction
Federalism may be defined as a system in which a central government and a
number of decentralised units, which are to some degree autonomous,
cooperate to some extent to attain common goals. Though some may prefer a
stricter definition, this one is convenient for the treatment of fiscal competition.
First, it includes not only countries with federal constitutions, but also those
where public service provision and taxation are decentralised, as well as
sovereign units forming an economic union. Second, it makes clear that fiscal
competition is an event related to one of the extremes of the continuum of
possible federal arrangements, namely when autonomy is fully exerted and no
coordination exists among the units.
1

Inter-jurisdictional competition may be passive, in the sense that independent
actions do not intentionally influence conditions faced by the unit or by other
jurisdictions. Or it may be active, meaning that tax or expenditure is
deliberately used as an instrument to pursue a given goal. Fiscal competition
may be horizontal, involving governments at the same level, or vertical,
involving competition between higher and lower levels of government. In
either case one cannot presume that fiscal competition is either welfare
enhancing or harmful.
This is the main question addressed by the vast and fast-growing literature on
fiscal competition originating from a seminal article by Tiebout (1956), and
Published by the Forum of Federations www.forumfed.org Publi par le Forum des fdrations

683
from Oatess systematisation of the existing economic theory on federalism in
the early 1970s (Oates, 1972). There is no simple answer. The general
inference is that the answer depends on several issues, prominent among
them the objectives of competing governments; what they are competing for;
how they compete; the behaviour of economic agents, especially their mobility
in response to fiscal stimuli; and the characteristics of the economic
environment, particularly the possibility of inter-jurisdictional externalities
arising from government actions.
This article does not set out to be a comprehensive survey of the extensive
literature on fiscal competition, but rather to extract from it typologies, and
some analyses and results that may help to organise a debate on the subject.
Therefore, this paper deals firstly with the objects and instruments of fiscal
competition. Secondly it presents some empirical evidence on the existence of
fiscal competition, and on the reaction of economic agents to inter-
jurisdictional differences in tax burdens and benefits from public spending.
Thirdly it considers the main tools that may be used to avoid or counteract
possible harmful effects of fiscal competition. Finally it summarises the
argument presented and speculates on the effect of globalisation on the roles
of central and decentralised government units.
2. How to compete, for what?
The traditional theory of fiscal federalism discusses the assignment of
economic functions of the public sector allocation, distribution and
stabilisation to different levels of government. The general conclusion is that
central governments should be responsible for macroeconomic stabilisation

684
and income redistribution, as well as for the provision of national public goods,
i.e. those for which the benefit area is the whole country (or economic union).
2

The economic case for decentralised governments rests on the existence of
public goods, the benefits from which are limited to a specific area or subset
of the population (local public goods).
The financing of local public goods in federations comes mainly from three
sources: taxes assigned to lower-level governments, inter-governmental
grants, and debt. Grants are inherently cooperative instruments, which if well
designed can serve several different objectives in a federation. Decentralised
taxation on the other hand, unless some degree of harmonisation exists, is
independently exerted and may distort resource allocation when economic
agents are mobile. To avoid distortions, theory recommends that only benefit
taxation should be applied to potentially mobile tax bases. But in the real
world, non-benefit taxation is the norm, being frequently used as an
instrument of active governmental competition. According to its object, fiscal
competition may be classified in three categories. First, decentralised units
compete in the provision of a bundle of public goods and services, trying to
improve their quality, reduce their cost and adjust supply to match residents
preferences. Second, they compete for funds to finance the provision of public
goods at the lowest possible tax price for residents. And third, competition
may have business investment as its object, to increase production,
employment levels and income within the unit.
2.1. Competition in the provision of public goods
Competition in the provision of public goods is the subject of the original
Tiebout model (Tiebout, 1956), as well as of more recent and richer models

685
(Oates and Schwab, 1988), which conclude, under a set of strong
assumptions, that this kind of competition is efficiency improving. In brief,
uncoordinated decision making would result in the provision of a variety of
fiscal packages (a bundle of public goods plus a tax price), so that mobile
individuals (or firms) could enjoy their preferred package by choosing to reside
in the locality where it is provided (voting with the feet). Competition is also
said to promote innovation in the provision of public goods and their diffusion,
and by benchmarking with other governments, to minimise organisational
costs of the public sector and reinforce accountability. Shah (2001) reports
that in Chile and Canada, school financing mechanisms encourage informal
benchmarking by citizens to guide their choice of schools.
Models that relax some of the strong assumptions mentioned in the preceding
paragraph show the reverse side of the coin. For instance, models employing
game theory drop the assumption that there is no strategic interaction in
response to policies of neighbouring jurisdictions, and find outcomes that
involve sub-optimal levels of public outputs (Wildasin, 1988).
When strategic behaviour exists, competition may stimulate the under-
provision of merit goods and social policies. In Brazil, for example,
municipalities are responsible for a large share of expenditures in public
health, financed partly by earmarked federal block grants and partly by their
own revenues. In metropolitan areas, individuals commute frequently across
cities, and since eligibility for public health services is not attached to
residence, municipalities providing better quality services are prone to attract
clients from surrounding cities. In fact, Ferreira (2002) found that
municipalities neighbouring the city of Rio de Janeiro spend less than the

686
expected value in public health services, both in per capita terms and as a
percentage of their respective tax revenues. In turn, the government of the city
of Rio de Janeiro did not seem to take into account the positive externalities
generated by its expenditures. The overall result is under-provision of public
health services in the metropolitan area.
3

Another interesting example is provided by the United States welfare system.
A 1996 reform decentralised welfare policy. States now have a large degree
of autonomy to decide on forms and levels of assistance to the poor.
However, if a state decides to increase its welfare benefits, it runs the risk of
attracting the poor from other localities. Their immigration increases state
welfare expenditures, but not the income tax revenue. To avoid becoming a
welfare magnet, and hence increasing the tax burden on the states better-
off residents, each state tends to reduce the value of the benefits provided. As
Brueckner (2000) points out:
because the concern about welfare migration depresses benefits in every
state, no state succeeds in repelling the poor by keeping its benefits low, and
each ends up being less generous than it would have been in the absence of
migration.
This reasoning points to a downward bias in the value of welfare benefits
under current institutional arrangements.
Oates (1999) recognised the shortcomings of decentralised systems in terms
of relief to the poor, but argued that a decision was made to accept the
downward bias as a price to be paid for the possibility of abandoning

687
unsatisfactory federal welfare programs, and looking for superior policy
alternatives. He asserted:
in a setting of imperfect information with learning-by-doing, there are potential
gains from experimentation with a variety of policies for addressing social and
economic problems. And a federal system may offer some real opportunities
for encouraging such experimentation and thereby promoting technical
progress in public policy.
He called policy experimentation in decentralised units laboratory federalism.
2.2. Competition for funds
The second category of fiscal competition is competition for funds to finance
the provision of public goods to residents at the lowest possible tax price. This
includes policies that aim to enlarge tax bases (or revenues), as well as
disputes for usually scarce costless or low-cost funds, provided by a higher
level of government.
Where the personal income tax is assigned to sub-national governments,
these units may attract the wealthy from other jurisdictions by reducing tax
rates or by providing a package of public goods tailored to their taste. Insofar
as pure (or almost pure) public goods are provided and therefore additional
consumers do not imply increase in the total cost of production newcomers
reduce the tax bill of the other residents. This beggar-thy-neighbour policy, if
successful, would imply higher tax prices for public goods elsewhere, and
therefore their under-provision. It might also weaken the power of income
redistribution policies. On the other hand, fiscally induced mobility may result

688
in a more homogeneous population in each jurisdiction, and lead to a closer
match between provision of and demand for local public goods.
Switzerland offers the best conditions for undertaking empirical analysis of
these points. Though there is a small federal income tax, cantons have the
basic power to tax income and wealth, while local jurisdictions levy property
taxes and a surcharge on cantonal direct taxes. Public spending is very
decentralised, and social assistance is a concern only of local and cantonal
governments.
4

Feld and Kirchgssner (2000) addressed their work to the question of whether
fiscal competition exists and what its effects might be. They concluded that
there is competition both among cantons and among cities; that taxes are
more important instruments than social transfers; and that tax competition is
stronger at local than at cantonal level. High-income earners choice of where
to reside depends on the amount of income tax they have to pay. Self-
employed earners are more responsive to tax stimulus than dependent
employees or retirees. For this last group, provision of public services plays a
more important role than taxation in making residence decisions. Feld and
Kirchgssner could not find any evidence that homogenisation of the
population brought any efficiency improvement. Fiscal competition on the
other hand, has not harmed decentralised income redistribution.
When origin-based commodity taxes are used, a jurisdiction may attract
consumers as opposed to residents by setting its tax rate below that of
neighbouring units. In this case, residents of higher tax areas can escape
taxation by incurring the transportation cost necessary to purchase certain

689
private goods in the low-tax jurisdiction. They will do so whenever the tax
differential exceeds the extra cost incurred.
Though cross-border trade and distance selling have always posed a problem
to tax designers, the recent expansion of e-commerce has made the need for
a solution more pressing. One such solution is the adoption of destination-
based commodity taxation. In this case, cross-border shopping and e-
commerce would compete in equal conditions with local retailers sales.
Nonetheless, destination principle schemes are administratively difficult to
implement.
5

When business, capital income or property taxes are in force in decentralised
government units, depending on economic conditions, tax exporting may
occur. Income and property taxes may be exported to foreign owners of
domestic companies or land. Business taxes may be shifted to residents of
other jurisdictions who consume the goods, through increases in the prices of
local output. Tax shifting is more likely when a locality produces a highly
specialised commodity like natural resources or tourist attractions. When tax
exporting occurs, residents of a particular area do not bear the full cost of the
public goods provided by the local government. This may give rise to
inefficient over-provision of these goods.
Brazilian municipalities provide a case in which tax exporting is preceded by
tax-base importing. These units levy an origin-based tax on services. The tax
base is determined nationally by means of a list of taxable services; and
municipalities are autonomous to set the tax rate. Most units charge a rate of
(or near) 5%; but some, which in normal conditions would have almost no tax

690
base, charge a lower rate of 2% or less, in order to attract tax base. Note that,
in contrast to the type of fiscal competition to be considered later on, the lower
rate does not attract investment or production to the territory of the unit, but
only the fiscal residence of the firm. To qualify as a resident, all a firm needs,
besides a signboard, is a rented room with a chair, a table, a telephone and
an attendant, costs which may be shared with several other firms. After the
tax base is imported, business continues to take place elsewhere; but the tax
on the services there rendered and consumed is paid to the municipality
where the headquarters is located.
Vertical tax competition may provide additional revenue to a sub-national
government at no extra cost for its constituency. This can happen whenever
central and decentralised units impose a tax on the same tax base, and the
lower-level tax may be credited against federal tax liability. If the
compensation takes the form of a deduction from the federal tax base, there
will be some increase in the overall burden faced by the taxpayer. This may
result in a reduction of the tax base available to both units, amplifying the loss
of revenue for the central government and reducing the gain of the
decentralised unit.
As noted by Wilson (1999), the negative externality imposed by the sub-
national unit reduction of the tax base does not necessarily imply that
taxes are inefficiently high in the new equilibrium. Under certain conditions,
the federal government may use its policy instruments to partially offset
inefficiencies at the sub-national level, or in some cases to achieve even an
efficient equilibrium.

691
Another form of lowering residents payments for public services is to compete
for access to funds provided at low or no cost by higher levels of government.
Shah (2001) notes that these funds are often allocated by programs, the
objectives of which are vaguely specified, and which lack focus on service
delivery and accountability to residents. This may give rise to pork-barrel
politics and waste. He illustrates his arguments with examples from Brazil and
Pakistan, where the president (prime minister) directed a substantial parcel of
disposable resources to his (her) home state (district), and from South Africa,
where provinces strategically overspent in local functions. They then claimed
they had no funds to provide the national functions, such as health and
education, which they administer. Of course, this is not to say that inter-
governmental transfers are undesirable. On the contrary, well-defined grants
play important roles in federal systems, including that of counteracting the
possible ill effects of fiscal competition.
2.3. Competition for business investment
Fiscal competition may aim to attract business investment to increase
production, employment levels and income within the jurisdiction. Passive
competition with the use of different non-benefit tax burdens being
explained, for example, by differences in tastes may lead to the same or the
opposite result. Instruments of this type of competition may be the tax
structure, the expenditures mix and regulatory policies, as well as tax
incentives and public services provided to specific firms.
Tax competition through lowering corporate income tax has been one of the
major fiscal issues in the European Union (EU) for many years. Those who
fear that fiscal competition will bring taxes on capital income to unduly low

692
levels, claim some degree of tax coordination. Another significant group takes
the opposite view that tax competition is welfare improving, and therefore that
corporate income tax should not be harmonised.
A recent paper (Zodrow, 2001) provides an overview of what economic
literature has to say in support of each of these opposite views. Zodrow starts
from a basic model (Zodrow and Mieszkowsky, 1986) which, under a set of
assumptions, concludes that tax competition leads to an inefficiently low level
of public services in all jurisdictions. Next, he reviews a wide variety of
extensions of the basic model that alter one or more of its assumptions. The
results are mixed: some identify potential gains, and others losses, from tax
competition; little is said about their magnitude. Coupling these results with
the observable reluctance of countries to give up their fiscal sovereignty, and
with the fact that some countries would be net losers from tax harmonisation,
Zodrow concludes that the case for it is tenuous. He suggests that modest
initiatives, like the Code of Conduct on Business Taxation (European
Commission, 1997), instead of attempts at full harmonisation of the income
tax, should be preferred.
However, unfettered tax competition in the EU ushers in a concern about the
future of re-distributive policies in the area (Sinn, 1994; Oates, 2001). Though
redistribution should preferably be assigned to central governments, the
European Community budget is too small to provide such programs, and there
is no intention to enlarge it significantly in the future. Therefore, each member
of the EU will have to support its own programs. The contention is that
increased factor mobility in the EU, in the absence of income tax coordination,
will force countries to rely more heavily on benefit taxation (which rules out

693
redistribution programs), or to incur a significant cost in terms of economic
growth, by taxing mobile factors to finance such programs. Furthermore, as
capital supply is generally more price-elastic than the labour supply, and
skilled labour is more mobile than unskilled, it may be expected that a wage
tax will fall more heavily on unskilled than on skilled labour, and that taxation
of capital income will be low, resulting in a more regressive tax system.
Another interesting question, addressed by Keen and Merchant (1997),
concerns the composition rather than the level of public spending in the
context of fiscal competition. They divide public spending into two categories:
utility enhancing (either public goods which complement private consumption
like recreational facilities, or re-distributional payments to some poorer
groups), and production improving (public inputs such as infrastructure or
general training). Since they assume that citizens are immobile and firms are
mobile, their conclusion is quite intuitive. In their own words, fiscal
competition leads to too many business centres and airports but not enough
parks or libraries.
Has this trend been observed in federal systems? And what is the impact of
expenditure competition among countries in a world with increasing mobility of
capital across borders? Those are open questions. But Keen and Merchants
result suggests that there is a case for coordination not only of taxes on
mobile bases, but also of domestic public expenditures.
A case may also exist for coordination of regulatory policies. The purpose of
regulation is to remedy market failures such as externalities and monopolistic
power. But it can affect a jurisdictions competitiveness. In particular, if profit-

694
maximising manufacturers take into account the compliance costs of local
regulation, governments may use lax regulation to attract business to the
territory of the unit. For example, there is a flow of literature on fiscal
competition that looks at the impact of environmental regulations on business
location (Levinson, 1996). In addition, the design of financial regulation can
potentially be used as an instrument to attract portfolio investment. Since
banking regulations are usually set at federal level, such competition generally
takes place among sovereign governments.
Instead of lowering taxes in an attempt to attract business, decentralised
governments may resort to concession of tax incentives, subsidies, and
provision of public inputs to specific firms. These are typical regional
development policy instruments. When used for decentralised industrial
policies, they may bring about destructive competition. The so-called fiscal
war among Brazilian states illustrates this point.
The practice of reducing state value added taxes (VATs) to attract investment
has been unlawful in Brazil since 1975, except in cases in which the intended
reduction is unanimously approved by the 26 states and the Federal District.
Yet the law has been disregarded, and tax competition among Brazilian states
has intensified since the beginning of the nineties. Foremost, in many cases,
is the competition for the wave of new automotive vehicle industrial plants that
have looked for a location in the country since 1995.
6

From the standpoint of any particular state, granting fiscal incentives to attract
investment seems worthwhile. Unless the beneficiary would choose to locate
his business in the state even in the absence of the incentive, the tax revenue

695
forgone would not exist anyway. In addition, aside from their direct impact on
production and employment, newly attracted firms induce additional economic
activity, creating still more jobs and income, and, of course, some tax
revenue.
If this were the whole story, state tax incentive would be a valuable
development tool. But when other states replicate the successful experience
of one, a destructive tax competition starts.
As the practice of granting incentives spreads, its efficacy fades. Since taxes
have been equally reduced everywhere, the fiscal benefit ultimately loses its
power to induce relocation of production. But revenue goes down in all states.
When the process reaches this stage, firms choose their location considering
only market and production conditions.
Pressed by greater spending and reduced tax collection, the less developed,
financially weaker states become unable to provide the services and public
works necessary to attract new business. In the final stages of the fiscal war,
the more developed states win all battles. Disparities already very large in
the case of Brazil naturally tend to increase.
The fiscal cost of the tax war for the country is very high. A recent dissertation
analysing three cases of newly installed vehicle factories (Silva, 2001),
concludes that in two cases the present value of the flow of subsidies exceeds
the value of private investment, and that the fiscal cost of creating a job is
over US$ 350, 000.
Furthermore, this does not seem to be a cost incurred to attract investment to
the country. The plants would probably be located in Brazil even in the

696
absence of the tax break.
7
Rather, this is the cost of attracting the investment
to one particular location within the country that, if the incentive had been truly
effective, would not be the one recommended by efficiency considerations.
An implicit assumption of most of the preceding discussion is that
governments are benevolent, i.e. they act in the best interests of the residents
of the jurisdiction. Public choice literature contends that a more realistic
assumption is that government officials and politicians have their own
objectives, act in their own interest or serve the interests of powerful interest
groups. In both these cases, instead of maximising the welfare of the
population, they will seek to maximise the size of the government budget.
Under this assumption, tax competition has the welfare-improving role of
counteracting Leviathan state tendency to over-expansion. In this context,
harmonisation of tax policies would serve the interest of the bureaucrats,
assuring monopoly of power to keep government revenue higher than it would
otherwise be.
8

3. Some empirical evidence
Do firms and individuals, as beneficiaries of welfare programs, consumers of
public and private goods, or factory owners, respond to fiscal stimuli? This is
an important question. A negative answer would mean that one should not
expect benefits or worry about the costs of fiscal competition. Though there is
substantial theoretical material on how economic agents react to tax and
expenditure differences across jurisdictions, there is not much empirical work
strictly related to the elasticity of tax bases in relation to observed differences
in the pattern of public spending or taxation.

697
One extensively debated question in the literature on expenditure competition
is whether there is welfare-motivated migration, i.e. whether or not welfare
recipients move from low to high-benefit jurisdictions. Considering the case of
welfare migration across US states, six out of a sample of eight studies found
evidence of migration, though two of them concluded that its magnitude is
small.
9
By contrast, Walker (1994), and Levine and Zimmerman (1995) could
not detect any evidence of welfare migration. Hence, the evidence is
moderately in favour of the hypothesis that migration exists, which may
indicate that states in the United States are in a non-cooperative equilibrium,
under-providing relief to their poor compared to what would be the optimal
outcome. However, the sensitivity of migration to welfare benefits is not high.
Meyer (1998), for example, found that a US$1000 increase in the annual
welfare benefit raises migration of single women to a region by only 6% over a
five-year period.
10

Instead of reducing overall welfare spending, states may protect themselves
from in-migration of the poor by limiting access to public goods. This is
generally done by restricting the status of residence. One can say that welfare
spending becomes a club good, since it is possible to exclude some
individuals from its consumption. Such action reduces incentives for the poor
to move, and as a consequence should lead to higher welfare transfers
compared to a situation without any exclusion. Evidence of such restrictions is
common in the history of the United States.
11
The existence of eligibility
conditions may partly explain why empirical studies do not find more
significant effects of welfare benefit differences on migration of the poor.

698
The studies of Figlio et al. (1997) and Saavedra (1998), instead of looking at
the responses to differences in welfare transfers, test the existence of
strategic interdependence between different states directly.
12
They provide
strong evidence that benefit levels in nearby states affect a given states
choice of benefit level.
Turning to tax competition, since there are not many federal systems where
sub-national governments have great freedom to set tax rates, the existing
evidence pertains to a few countries. Countries belonging to the Organisation
for Economic Cooperation and Development (OECD) for which there is
enough data available are the United States, Canada, Germany and
Switzerland. Even in Germany, the local taxing autonomy is mainly confined
to the business tax. Most of the literature does not test the existence of tax
competition, but the sensitivity of a given tax base to the level of the tax rate.
There is a vast set of empirical studies looking at the impact of capital taxes
on several different measures of business activity. Most of the studies are
applied to the United States, studying the impacts of differences in income tax
rate across states, and differences in property taxation within a given state.
Table 1 summarises the results found in a survey of the United States
(Wasylenko, 1997). The cells of the table report the number of studies where
an elasticity measure was estimated, the number of those studies in which the
tax elasticity was statistically significant (in parenthesis), the range of elasticity
estimates (in brackets), and the median elasticity.
Estimates in Table 1 indicate the percentage decrease in the dependent
variable when the tax rate in a given location is 1% higher than in a nearby

699
location. For example, the impact of a business tax 1% higher in a given state,
compared to other states, is a 0.11% decrease in employment in that state
(column 2, line 1). The main conclusions are summarised below.
Estimates of response to tax differentials vary widely. Intra-regional
differences in tax rate have a larger impact on business location than
differences across states (or inter-regional). The inference is that once the
locality is chosen (taking into account a set of regional attributes like
agglomeration, cost of labour, size of the market, quality of education,
infrastructure of transportation etc), the specific location (in which
neighbourhood or suburb) will be strongly determined by tax
considerations.
13

Table 1 Summary of econometric results of tax effects on business location
Inter-regional or
Inter-state
Studies
Intra-regional Studies
Dependent Variable
Overall Tax Elasticity
Business Tax
Elasticity
Property or
Business Tax
Elasticity
Aggregate Data
Total Employment 6 studies (5)
[-0.85, 0]
-0.58
3 studies (2)
[-0.16, 0]
-0.11
4 studies (3)
[-1.95, -0.81]
-1.85
Manufacturing Employment 13 studies (8)
[-1.54, 0.05]
-0.10
2 studies (1)
[-0.26, 0]
1 study (1)
-0.79
Investment in Manufacturing 6 studies (3)
[-1.02, 0.54]
-0.60 or 0
7studies (6)
[-0.36, -0.10]
-0.20

Gross State Product, Income
or Value Added
12 studies (7)
[-0.88, 0.27]
-0.07
1 study (0)
-0.14

Micro Data

700
Manufacturing Plant Berths or
Location
3 studies (2)
[-0.40, 0]
-0.18
19 studies (15)
[-15.7, 0.6]
-0.20
5 studies (4)
-2.70, 0.62]
-1.59

The wide range of the elasticity estimates has less to do with the type
of activity being measured than with the variations in data, time
periods, and other variables used in the estimation equation. In effect,
the results change depending on which variables are included in the
estimation equation, or which time period is analysed.
In particular, adding controls for the type and level of public goods
supplied by each location significantly affects the econometric results.
Business-friendly regulations and public spending that enhances
productivity enable a given location to set a higher capital and/or
property tax rate. In other words, local attributes increase the tax
setting power of a given jurisdiction and should be taken into account
when estimating business responses to tax differences.
In a different vein, some studies test the existence of strategic
complementarity on business tax setting between jurisdictions. Ladd (1992)
found statistical support for the hypothesis that neighbouring jurisdictions
mimic each others tax policy in the United States. Bttner (1999) tested the
existence of tax competition from the relationships between the levels of
capital income tax rate in German districts. Like Ladd (1992), he found
evidence that tax rates are positively related to neighbours tax rates.
14

Some evidence on cross-border shopping may also be found in the empirical
literature on tax competition. Using data from the United States, Due and

701
Mikesell (1994) find that a 1% differential in sales taxes results in a shift from
1% to 6% of purchases from higher to lower tax areas. In Canada, a study of
this phenomenon found little evidence of cross-border trade in the Ottawa-Hull
area in the 1970s, when the inter-provincial rate differential was 3% (Dufour
and Vaillancourt, 1982).
In Europe, some regimes of tax exemption for non-residents lead to cross-
country shopping. For example, in the Schleswig-Holstein border between
Denmark and Germany, Germans used to buy cars in Denmark while Danes
bought spare parts in Germany. This was a consequence of regulation and
taxation, which led to completely different final prices in the two countries
(Economist, 29.11.2001).
4. How to cope with fiscal competition
One of the challenges facing areas (countries, unions or even the whole
world) where fiscal competition develops is how to reduce the welfare loss
from its many facets without sacrificing the benefits of decentralisation.
A country may impose restrictions on beggar-thy-neighbour policies by means
of a constitutional provision or national laws binding the decentralised units.
However, such restrictions may be difficult to enforce. Authorities would have
to keep track of a large assortment of fiscal instruments, including disguised
ones. It would be difficult to tell whether these instruments were directed to
competition, or to other objectives that they can also serve. And long judicial
battles might be necessary to determine whether or not the act of one
decentralised unit caused any damage to the affairs of another. Besides, there

702
is a risk of putting welfare-enhancing competition in the same bag, and
preventing it as well.
A high degree of centralisation of taxing powers, coupled either with transfers
to decentralised units (as is the case of Argentina), or the assignment of tax
legislation to the federal government (as in Germany), are possible solutions.
They have in common the disadvantage of eliminating one important facet of
federalism, namely the autonomous determination of the size of each sub-
national units budget. Vertical coordination (tax collection agreements, tax-
base sharing, abatement of sub-national from federal taxes), which is
extensively used in Canada, results in more uniform tax bases, leaving space
for decentralised decisions on the size of the budget, but also for some
competition.
Inter-governmental transfer mechanisms can be designed to reduce the
detrimental effects of fiscal competition without sacrificing the benefits of
decentralisation. The theory of fiscal competition is concerned with the
existence of externalities generated by the action of a given jurisdiction over
the residents of another, and with the consequences when tax and
expenditure decisions do not take such externalities into account. Economic
theory prescribes the use of a system of inter-jurisdictional transfers whereby
a given unit pays taxes for the negative spillovers, and receives subsidies for
the positive spillovers that it promotes. Such Pigouvian transfer systems
would theoretically drive the system to an efficient decentralised equilibrium
(Varian, 1992). Unfortunately, implementation of this ideal transfer scheme is
impossible, and federal countries use non-optimal schemes.

703
In the case of expenditure competition, the under-provision of transfers to the
poor resulting from decentralisation may be partially offset by earmarked
grants from the central to sub-national governments. This is the case with the
decentralised provision of public education and health in Brazil, as well as that
of US states direct assistance to families below poverty line.
Earmarked transfers may either take the form of block grants or matching
grants. Under the block grant each jurisdiction receives a lump sum from the
central government, the magnitude of which is independent of the level of
jurisdiction contribution to the provision of the public good. Under the
matching grant, individual jurisdictions determine the level of expenditure, and
the central government pays a fixed share of a jurisdictions total outlay. The
theory of expenditure competition prescribes a matching-grant system
because it reduces the marginal cost (faced by the states) of providing welfare
programs, leading to a higher equilibrium level of expenditure. Under the block
grant system, states tend to spend only the amount of the lump-sum transfer
coming from the central government.
15

Harmonisation of fiscal policies may also be used as a tool to reduce the
negative effects of fiscal competition, while preserving the advantages of
decentralised policies. In the case of unions where central governments
have a very small budget and decentralised units are sovereign jurisdictions
that cannot be legally bound except by voluntary subscription to a treaty,
harmonisation may well be the only feasible instrument to cope with fiscal
competitions undesirable effects.

704
As mentioned in the introduction, fiscal competition is an extreme case in
which members of a federation act independently, without any scope for
cooperation. Harmonisation is a move to a position in which some cooperation
exists. This may range from token coordination, which is presently the position
with respect to EU corporate income taxes, to full integration, a position in
which the units give up their autonomy or sovereignty, as is the case with the
monetary policies of countries in the European Monetary Union.
Much has been said about the need for harmonisation of fiscal policies among
the European countries as they engage in deeper integration. And much has
been said against harmonisation, particularly by those who believe in
Leviathan. But even discarding the hypothesis that harmonisation will be the
instrument to assure large-scale government, it must be recognised that the
implementation of such a coordination scheme is far from trivial, especially in
the economic union.
Firstly, a contract among sovereign countries must consider a wide range of
possible non-cooperative strategies that should be ruled out. It is probably
impossible to cover every alternative. For example, harmonisation of the tax
structure may be put at risk by lenient enforcement in a given jurisdiction.
Secondly, when dealing with sovereign countries, such a federalist pact is
not enforceable if one party decides to act uncooperatively. Hence, an
organism to supervise and enforce the agreement must be built before such a
contract is designed. The question is: are EU members prepared to give up
their fiscal sovereignty? This is a sine qua non condition for deepening the
harmonisation process.

705
The answer to this question is dependent on a number of factors, important
among them the answer to another question: how significant are the gains to
be reaped from tax coordination? There are few answers to this question in
economic literature, and most of them are provided in the context of highly
simplified models.
A recent paper (Sorensen, 2001) developed a tax competition model that
relaxes many of the restrictive assumptions of previous modelling efforts, in
an attempt to provide more reliable guidance to policy makers. Sorensen uses
the model to offer quantitative estimates of the welfare gains from tax
coordination. He considers the example of global coordination, whereby all
countries worldwide coordinate their tax policies, and of regional coordination
whereby only a subset of countries (the union) coordinates their policies. His
main conclusions are:
that the gain from regional tax coordination is only a small fraction of the
potential gain from global coordination if capital mobility is perfect. With
imperfect capital mobility between the tax union and the rest of the world,
there is greater scope for regional tax coordination, although the welfare gain
will almost certainly be well below 1% of Gross Domestic Product (GDP) and
will accrue mainly to countries with high initial capital income tax rates.
In short, the reward for surrendering fiscal sovereignty seems to be too low.
5. Summary and a note on globalisation
Fiscal competition is a natural companion of decentralisation. Potentially it
always exists, since it is the consequence of differences among jurisdictions,

706
and not necessarily of intentionally promoted discrepancies; there are no two
identical government units in the world. Practically, the manifestation of fiscal
competition depends on the intensity of the divergences, and on the reaction
of the economic entities in face of the array of options offered by
decentralisation.
Fiscal competition takes several forms, uses diverse instruments and may
bring about a number of different outcomes. An impressive amount of
theoretical work tries to model the phenomenon. Overall the results are quite
sensitive to the set of assumptions adopted in the analysis. Therefore there
are results to almost all tastes. The state of the art, as expressed by Wilson
(1999), is that competition among governments is now seen as a less
straightforward phenomenon than perhaps originally envisioned. And of
course, there is space for further modelling, with the introduction of complexity
that may bring the ideal closer to the real world.
Proving the practical existence of fiscal competition, and verifying its impact
on factors of production and consumer movement across jurisdictions is an
important step. Knowledge about the effects of competition on the economic
agents, and on the intensity of their reaction to the fiscal stimulus, is helpful for
the conception of mechanisms to curb or invigorate government competition,
whichever is the case. But which is the case?
The results given by what may be called the traditional tax competition models
show that tax competition tends to distort the allocation of resources,
promoting welfare losses. Given that these losses exist, they should be
weighed against the possible concomitant gains from expenditure competition

707
(e.g. ideal environment for public policy innovations, and a closer match
between public goods provision and local preferences). The existing literature
provides almost no evidence about the magnitude of these gains and losses,
and further research on this difficult empirical problem is necessary to fill this
fundamental gap.
In the absence of clear-cut conclusions from either theoretical or empirical
literature, the wisest attitude toward fiscal competition seems to be to avoid
extreme measures either to impede or to enhance competition. Hence,
controls or re-centralisation may be welfare-reducing measures insofar as
they eliminate political competition among jurisdictions or create the
environment for the Leviathan to rise. Of course the best course of action is,
whenever possible, to adopt measures that reduce welfare losses without
sacrificing the benefits of decentralisation. Carefully designed inter-
governmental transfers and cautiously conducted harmonisation processes
seem to be the most promising instruments.
Finally, some conjectures should be made on the impact of the globalisation
of economic activities on fiscal competition. Globalisation and regional
integration restrain fiscal sovereignty, insofar as factor mobility and growing
trade-flows require that domestic policies, including taxation, follow
international patterns more closely. Homogenisation of central governments
practices may induce decentralised units to assume the task of attracting
foreign direct investment by increasing the provision of local public inputs.
Furthermore, international competitiveness is increasingly contingent on the
existence of skilled labour, which depends on education and training outlays
that are typically decentralised government functions.

708
Therefore, it should be expected that the intensity of fiscal competition will
increase in the near future, and that sub-national units will be competing not
only among themselves, but also in the world market. They will probably
bypass the national governments and negotiate directly with firms regarding
the location of their business.
Given that skilled labour, infrastructure and other local public inputs are
tokens in these negotiations, the less developed regions of a country, or
indeed the world, will be at a disadvantage. Regional disparities (as well as
personal income concentration) will tend to increase, which suggests that
central governments and international institutions should amplify their
personal and regional re-distributive efforts in order to neutralise this
undesirable trend.
References
Blank, R.M., 1998. The Effect of Welfare and Wage Levels on the Location
Decisions of Female-Headed Households. Journal of Urban Economics, 24,
186-211.
Besley, T. and Case, A., 1995. Incumbent Behavior: Vote-Seeking, Tax
Setting, and Yardstick Competition. American Economic Review, 85(1) March,
25-45.
Bird, R.M. and Gendron, P.P., 2000. CVAT, VIVAT, and Dual VAT: Vertical
Sharing and Interstate Trade. International Tax and Public Finance, 7 (6),
753-761.
Borjas, G.J., 1997. Immigration and Welfare Magnets. Harvard University,
mimeo.

709
Brueckner, J.K., 2000. Welfare Reform and the Race to the Bottom: Theory
and Evidence. University of Illinois Urbana-Champaign, mimeo.
Bttner, T., 1999. Determinants of Tax Rates in Local Capital Income
Taxation: A Theoretical Model and Evidence from Germany. CESifo Working
Paper Series No. 194, September 1999.
Due, J. and Mikesell, J.L., 1994. Sales Taxation: State and Local Structure
and Administration. Washington: Urban Institute Press.
Dufour, J-M. and Vaillancourt, F., 1982. Provincial and Federal Sales Taxes:
Evidence of Their Effect and Prospects for Change. In: W.R. Thirsk and J.
Whalley, eds. Tax Policy Options in the 1980s. Montreal: Canadian Tax
Foundation.
Edwards, J. and Keen, M., 1996. Tax Competition and Leviathan. European
Economic Review, 40, 113-134.
Enchautegui, M.E., 1997. Welfare Payments and Other Determinants of
Female Migration. Journal of Labor Economics, 15, 529-554.
European Commission, 1997. Toward Tax Co-ordination in the European
Union: A Package to Tackle Harmful Tax Competition. COM (97) 495 final.
Feld, L. and Kirchgssner, G., 2000. Income Tax Competition at the State and
Local Level in Switzerland. CESifo Working Paper Series No. 238, January
2000.
Ferreira, S.G., 2002. Municpios: Despesa com Sade e Transferncias
Federais. Informe-SF No. 38. rea de Assuntos Fiscais e de
Emprego/BNDES.

710
Figlio, D.N., Van Kolpin, W. and Reid, W., 1999. Do States Play Welfare
Games? Journal of Urban Economics, 46 (3) November, 437-454.
Gordon, R.H. and Wilson, J.D., 2001. Expenditure Competition. National
Bureau of Economic Research Working Paper 8189, March 2001.
Keen, M., 2000, VIVAT, CVAT, and All That: New Forms of Value Added Tax
for Federal Systems. Canadian Tax Journal, 18(2), 409-424.
Keen, M. and Marchand, M., 1997. Fiscal Competition and the Pattern of
Public Spending. Journal of Public Economics, 66, 33-53.
Ladd, H., 1992. Mimicking of Local Tax Burdens among Neighboring
Counties. Public Finance Quarterly, 20(4), 450-467.
Levine, P.B. and Zimmerman, D.J. 1995, An Empirical Analysis of the Welfare
Magnet Debate Using the NLSY. National Bureau of Economic Research
Working Paper No. 5264.
Levinson, A., 1996. Environmental Regulations and Manufacturers Location
Choices: Evidence from the Census of Manufacturers. Journal of Public
Economics, 62, 5-29.
Locke, W. and Tassounyi, A., 1996. Local Tax Base Endogeneity and Tax
Competition: A Consideration of the Greater Toronto Area. Paper presented in
the Canadian Economics Association Annual Meeting, Brock University,
Ontario, 1996.
McLure, C.E. (Jr), 2000. Implementing Subnational Value Added Taxes on
Internal Trade: The Compensating VAT (CVAT). International Tax and Public
Finance, 7(6), 723-740.

711
Meyer, B., 1998. Do the Poor Migrate to Receive Higher Welfare Benefits?
Northwestern University, mimeo.
Mintz, J. and Tulkens, H., 1986. Commodity Tax Competition between
Member States of a Federation: Equilibrium and Efficiency. Journal of Public
Economics, 29, 133-172.
Mora, M. and Varsano, R., 2001. Fiscal Decentralization and Subnational
Fiscal Autonomy in Brazil: Some Facts from the Nineties, IPEA, Texto para.
Discusso No. 854, December 2001.
Oates, W., 1972. Fiscal Federalism. New York: Harcourt Brace.
Oates, W., 1999. An Essay on Fiscal Federalism. Journal of Economic
Literature, 37, 1120-1149.
Oates, W.E., 2001. Fiscal Competition and European Union: Contrasting
Perspectives. Regional Science and Urban Economics, 31, 133-145.
Oates, W.E. and Schwab, R., 1988. Economic Competition among
Jurisdictions: Efficiency-Enhancing or Distortion-Inducing? Journal of Public
Economics, 35, 333-354.
Poddar, S., 1999. Considerations in the Design of a VAT at the State Level in
India. Paper prepared for the VAT Seminar for Senior Officials from the
Centre and the States, July 1999.
Rauscher, M., 1998. Leviathan and Competition among Jurisdictions: The
Case of Benefit Taxation. Journal of Urban Economics, 44, 59-67.
Saavedra, L.A., 2000. A Model of Welfare Competition with Empirical
Evidence from AFDC. Journal of Urban Economics, 47 (2) March, 248-279.

712
Shah, A., 2001. Interregional Competition and Federal Cooperation To
Compete or to Cooperate? Thats not the Question. Paper presented at the
International Forum on Federalism in Mexico: Local and Global Challenges.
Vera Cruz, Mexico, November 2001.
Silva, M.A., 2001. Guerra Fiscal e Finanas Federativas no Brasil: o Caso do
Setor Automotivo. Universidade Estadual de Campinas, Masters dissertation.
Sinn, H-W., 1994. How Much Europe? Subsidiarity, Centralization and Fiscal
Competition. Scottish Journal of Political Economy, 41, 85-107.
Sinn, S., 1992. The Taming of Leviathan: Competition among Governments.
Constitutional Political Economy, 3, 177-196.
Slack, E., 1994. Non-Residential Property Taxation and Competitive
Advantage in the Greater Toronto Area. Canadian Urban Institute: Urban
Focus Series No.94.
Sorensen, P.B., 2001. International Tax Coordination: Regionalism versus
Globalism. CESifo Working Paper No. 483, May 2001.
Southwick, L. (Jr), 1981. Public Welfare Programs and Recipient Migration.
Growth and Change, 12, 22-32.
Tiebout, C., 1956. A Pure Theory of Local expenditures. Journal of Political
Economy, 64, 416-424.
Varian, H., 1992. Microeconomic Analysis. 3rd ed. New York: W.W. Norton &
Co.
Varsano, R., 2000. Subnational Taxation and the Treatment of Interstate
Trade in Brazil: Problems and a Proposed Solution. In: S.J. Burki et al., eds.

713
Decentralization and Accountability of the Public Sector. Proceedings of the
Annual Bank Conference on Development in Latin America and the
Caribbean. Washington, D. C.: World Bank, 339-356.
Walker, J.R., 1994. Migration among Low Income Households: Helping the
Witch Doctors Reach Consensus. Institute for Research on Poverty,
University of Wisconsin-Madison, mimeo.
Wasylenko, M., 1997. Taxation and Economic Development: The State of the
Economic Literature. New England Economic Review, March/April, 36-52.
Wildasin, D.E., 1988. Nash Equilibria in Models of Fiscal Competition. Journal
of Public Economics, 35, 229-240.
Wilson, J.D., 1999. Theories of Tax Competition. National Tax Journal, 52,
269-304.
Zodrow, G.R., 2001. Tax Competition and Tax Harmonization in the European
Union. Paper presented at the Conference on Tax Policy in the European
Union, The Hague, October 2001.
Zodrow, G.R. and Mieszkowski, P., 1986. Pigou, Tiebout, Property Taxation,
and the Under-provision of Local Public Goods. Journal of Urban Economics,
19, 356-370.

714

1
Fiscal competition occurs in a situation in which each federate unit independently decides a tax or expenditure
policy. This does not preclude concomitant cooperation in other fiscal and non-fiscal policies.
2
Typically, decentralised units do not have monetary policy instruments, and being highly open, are unable to
influence macroeconomic conditions using fiscal devices. Income redistribution policies on the other hand, are
constrained by the mobility of economic agents. If a tax-the-rich-benefit-the-poor policy were locally pursued, higher-
income households would tend to leave a jurisdiction and an inflow of poor families would be stimulated.
Notwithstanding, decentralised government units often perform functions financed by the wealthy or by all, where
benefits accrue mainly to the poor; local programs that provide cash or in-kind relief to the poor are not uncommon.
3
It is said as a joke but it is not too far from reality that Osasco, a municipality in the metropolitan area of So
Paulo, solved all its health problems by buying ambulances that remove its sick residents to hospitals in the city of
So Paulo.
4
It should be noted, however, that social assistance expenditures are a small fraction of total expenditures.
5
On the design of such schemes see Poddar, 1999; Varsano, 2000; Bird and Gendron, 2000; McLure, 2000; and
Keen, 2000.
6
This and the next few paragraphs on Brazilian states fiscal war draw from Mora and Varsano, 2001.
7
A possible but improbable alternative location, given that the market to serve is chiefly the Mercosur, would be
Argentina. If this alternative had, in fact, been considered and discarded because of the incentives, the fiscal cost
cannot be said to be in vain. But Brazilian state policy would be unduly inflicting a loss on the partner.
8
Some formal Leviathan-type models are presented in Sinn (1992), Edwards and Keen (1996), Rauscher (1998),
and Gordon and Wilson (2001).
9
Southwick, 1981; Blank, 1988; Borjas, 1997; Enchautegui, 1997; Meyer, 1998.
10
Most of these studies are based on the AFDC (Aid to Families with Dependent Children), in which money is given
to the single mother. Meyer (1998), for example, finds that single mothers migrate more readily in response to higher
welfare benefits than single women without children, who are not eligible for the benefit. This is additional evidence of
welfare migration.
11
Brueckner (2000) observes that some states imposed severe restrictions by denying any welfare benefits to poor
migrants over a waiting period as long as one year. Such restrictions were struck down by the Supreme Court in
1969, but states responded by instituting a two-tier benefit scheme, under which the benefits earned by migrants
during their waiting period corresponded to the benefit level in their state of origin. The most well known case is that
of Wisconsin, which protected itself against migration from Illinois, a traditionally less generous state.
12
In the presence of fiscal competition, one should expect strategic complementarity among governments. For
example, when the neighbour increases the tax rate on capital, the given state (or country, or municipality) will act in
the same direction, and vice versa.

715

13
At the intra-provincial level, two papers concerning tax competition in Canada should be mentioned. Locke and
Tassounyi (1996) found that business migrates from metropolitan Toronto to the vicinities which charge lower non-
residential property tax. Slack (1994), looking at data from Ontario, inferred that higher non-residential property taxes
may discourage businesses from location in a given municipality. She also concluded that property tax differentials
are not a major factor in the decision to locate in one metropolitan area or another; but once a metropolitan area is
chosen, they affect the decision about the specific municipality in which to locate.
14
The explanation for a positive correlation between tax rates in a neighbourhood may be a result of classic
competition, since the tax base is volatile. Alternatively, it might be a result of political competition. Voters compare
policies in the neighbouring district with those of their own district. The mayor does not get re-elected if his (or her)
policy happens to be worse than the one in the neighbouring district (Besley and Case, 1995).
15
The United States welfare reform of 1996 transited from a matching grant to a block grant system, and gave more
freedom to the states to define their own policies. Brueckner (1999) argues that this switch may cause a reduction in
welfare spending in the long run, which could only be corrected by going back to the matching grant mechanism.