ANALYSE OF TAX COMPETITION IN BRAZIL
USING THE SYSTEM DYNAMIC APPROACH
Abstract
The present study analyzes the systemic interaction of tax competition in Brazil,
aiming to identify the elements that perpetual this practice and its effects on the
economy. It should be noted that according to the proposed model, the results obtained
with the tax competition is not even optimal condition for the less favored resource
region (Northeast). Since, upon the incursion of specific investments in infrastructure
(in the same amount of tax waivers), the behavior of GDP, infrastructure and revenue
would be higher in this region. For this reason, it is evident that in the sub-national level
(decentralized) the only alternative industrial policy for the less affluent is the granting
of tax incentives. However, this type of policy is effective palliative and temporary and
does not constitute, per se, a sustainable policy to fix the route of economic
concentration. It remains for the central government to put private investment policies
in place to income the attractiveness of these regions. Otherwise, these regions do not
have incentives to reduce tax competition.
Key words: tax competition, macroeconomic, government investments
ANALYSE OF TAX COMPETITION IN BRAZIL
USING THE SYSTEM DYNAMIC APPROACH
1. Introduction
The industrialization, leveraged since 1930, developed a scheme of inter-
regional division of labor, setting the framework for regional disparities (DULCI,
2002). The interconnection of regional markets, previously isolated, together with the
adoption of policies to protect the domestic market and resulting economic dynamism
of coffee contributed to the industrial concentration in the Southeast, especially in Sao
Paulo. This industrial complex gained strength from the adoption of an exchange rate
policy directed at restricting imports, and increased the centrality of the state economy
at the expense of less developed regions.
Given the advanced process of uneven development, it became evident the need
to adopt measures to mitigate the distortion in the wide portion of land placed outside
the process of economic growth. In this sense, the federal government created the
Superintendence of Northeast Development (SUDENE) and Amazon Development
Superintendence (SUDAM). According Dulci (2002) these agencies consolidated line
of action of the federal government with regard to the efforts of economic recovery in
peripheral areas - in relation to the axis of the industrial south - through institutional
mechanisms, particularly in the tax field.
For the states with an intermediate development position, the delay did not get in
dimensions to justify the strategic action of the federal government adopted political
and institutional resources to reduce their economic backwardness. Given the political
importance of these states, their disputes for investment attraction resulted in tensions
between federal entities. Therefore, it was necessary the active role of the federal
government to balance competing interests and ensure national unity. In this sense
public initiatives have been promoted for expansion and modernization of infrastructure
and promotion of development (SILVA; OLIVEIRA, 2007).
The active participation of the federal government in correcting the distortion
was legitimized by the need to uphold the federal pact in heterogeneous conditions of
the Brazilian states. The presence of a strong and inductor state assumed centralizing
character, both in tax collection, and the application of resources. From mid-1980
through the fiscal crisis and the decline of the military regime, there was a trend and
democratic decentralization, which gave more power to the states and municipalities.
The thrust towards decentralization of political and financial support was given by the
1988 constitution, which stipulated that each state has autonomy to legislate and fix the
rates of taxes imposed by the state.
The decentralization promoted by the constitution of 1988 reduced the income of
the Union, which, before the economic crisis and the need for fiscal adjustment, adhered
to the logic of phased withdrawal of the "State" in the economy. This, coupled with the
progressive opening of the Brazilian economy, provided a favorable environment to the
state contests. Given that economic liberalization attracted an increasing flow of foreign
capital interested in investing in the country and that states had the freedom to fix their
tributes, each federative entity has to compete with others in attracting private
investment into their respective territories. The tools used in these disputes are the tax
exemptions and tax expenditures, and given the name "tax war" (Arbix, 2000).
Wildasin (2011, p. 1313) states that the competition for mobile factors of
production can be better understood when examined through a "Explicitly dynamic
framework", that is, through an analytical framework to present the dynamic causal
relationships occur where interactions. According to the author, this approach allows to
detect the speed and magnitude of adjustment of economic variables in response to
policy changes and to compare their results in different scenarios.
Therefore, this study seeks to understand the interaction of systemic emerging
tax competition for new investments in Brazil, and to assess its economic impact. The
analyze of this problem from the perspective of system dynamics possible to find
counter-intuitive elements that encompass the intergovernmental competition, and to
draw conclusions and point out possible ways that go beyond the prescriptive
requirements in the literature that focuses these theme.
Besides this introduction this study contained three sections. The next section
presents the theory fundaments of system dynamics, considered the archetypes, the
simulation model of tax competition between two regions, as well as the handling of the
model and data sources. Then we describe the simulation results and evaluation of
scenarios and finally, the last section presents the conclusions and suggestions for
further work.
2. Methodology
This study covers the Southeast and Northeast regions which account together
about 70% of Brazilian GDP. The choice of such regions is due to the fact that properly
represents the contrasts in terms of capital allocation and infrastructure resulting from
the process of uneven development. Arbix (2000) and Dulci (2002) also add that these
two regions are composed of some of the states that stood out as the great promoters of
the "war" on several fronts.
In order to meet the objectives of this work a systemic model of tax competition
for new investments between two regions using the system dynamic approach are
developed. To identify causal relationships and feedback loops two systemic
archetypes: "Success to the successful" and "escalation" are combined. The
methodology of system dynamics was initially developed by Forrester (1961), in order
to track the effects of policies on the state of the system. The method of system dynamic
allows the calibration of the behavioral parameters that are adherent to the reality of the
competitive dynamics between the analyzed regions. Thus, after calibration, is possible
to perform simulations and to analyze scenarios.
The process of tax competition is systemic in nature, since the adoption of one
implies downside the other. This individual behavior brings out an unintended path that
restricts the ability of regional governments to ensure the provision of public goods and
services, as well as the incursion of systematic regional development policies.
Present in several areas of knowledge, the phenomenon of emergence is
understood as a process of self-organization that results from the interaction between the
actors. The emergency occurs so often unintentional and the result of individual actions
which, although considered rational in the strict sense, can not anticipate the collective
result of these actions (Jervis, 1997). Bueno (2009) points out that the emergence is a
property of complex systems that exhibit dynamic complexity.
The dynamic complexity derives largely from the idea that the interaction
between systemic agents in the environment occurs in the form of feedback loops,
which may take the form of reinforcement (positive) or balance (negative). Sterman
(2000) puts this environment that apparently isolated actions of individuals can trigger
other reactions, changing, in the subsequent period, the conditions under which the
decisions were taken at first. Thus, individuals may also change their strategies in
response to the strategies of other agents, as well as changes in physical and institutional
environment. The result is that in dynamically complex systems, decisions produce, in
general, results not intended by their makers, where the effects are away from their
causes through delays (lags).
As a result, the system dynamics is the most appropriate approach to analyze the
dynamic interaction in complex systems, it consists of a set of techniques designed to
evaluate systems controlled by feedback loops. This approach enables the construction
of co-evolutionary models for institutional dynamics, and to identify the most relevant
causal chains present in the model.
2.1 Archetypes
Based upon the studies of interactions, the system dynamics group at the
Massachusetts Institute of Technology (MIT) identified eight common systemic
structures, which apply to a wider range of social interaction situations. These structures
are called systemic archetypes, being developed from different combinations of loops
and strengthening the balance sheet (Kim, 2000). The following are two of these
archetypes: the “escalation” and “success to success”.
The archetype "escalation" expresses the race between two players (regions) that
wants to achieve a privileged position in relation to your opponent. Thus, if the region
puts the ongoing decisions that improve their relative position (policies on tax
incentives), the region B considers this practice as a threat to their security. In response
to this threat, B also decides to grant tax incentives, rebuilding their relative position (in
terms of the attractiveness of investments). However, A also believes the initiative to B
as a threat and, therefore, responds with expansion of the policy initiated. Although the
individual action produces loops balance - that is, considering only their isolated action,
the increase in tax incentives improves its relative position and therefore discourages the
continuation of this expansion in a subsequent period - each agent reacts to the initiative
of another with intensifying process already underway, resulting in a process of
strengthening the fiscal war. The tax competition literature classifies this phenomenon
as "race-to-the-bottom" (ZODROW; MIESZKWSKI, 1986). Figure 1 represents the
causal diagram of the escalation.
Escalation
Quality of A's Position
Activity by A. Rosie WB's Activity by B
to ~s Ne to _#t
Figure | - Archetype "Escalation".
Source: Kim (2000)
The second archetype of success for the goods expressed a successful scenario
of hypersensitivity to initial conditions. In this situation the region that gets a head start
even if temporary, such as better infrastructure and logistics, as a result of this relative
gain, follows a path of gradual concentration of investments in its territory. If not offset
by other policies or loops, the loop reinforcement tends to produce increased industrial
concentration in the first region, while the other becomes increasingly poor. Figure 2
illustrates the said archetype.
Sucess to the Sucessful
Sucess of A Sucess of B.
= +
+ s
Allocation to A
RI instead of B R2
+
Resources to A- ‘Resources to B
Figure 2 - Archetype "Success to the Successful".
Source: Kim (2000)
2.2 Simulation model
The system model employed in this study was adapted from the models
proposed by Wilson and Wildasin (2004), Keen and Marchand (1997), Zodrow and
Mieszkwski (1986). As a result, the proposed model was built from six basic postulates,
which are:
1. There are only two types of jurisdiction in the country:
a) Rich region - with better allocation of capital, infrastructure and logistics.
Therefore more attractive to private investment, and
b) Poor region - with low allocation of capital, infrastructure and logistics. Therefore
less attractive to private investment;
2. Individuals are homogeneous and no mobility between regions (KEEN;
MARCHAND, 1997);
3. Capital is mobile only in the form of new investments, that is, once immobilized the
inversion is not possible to remove it to another jurisdiction;
4. Each sub-national government seeks to maximize the welfare of its citizens and has
autonomy in the conduct of fiscal policy, but should maintain the fiscal balance
(KEEN; MARCHAND, 1997, MINTZ, Tulkens, 1996, ZODROW; MIESZKWSKI,
1986);
5. There is no spatial spillover or externality tax, i.e. tax policy and public goods in
each jurisdiction does not affect the results of others (ZODROW; MIESZKWSKI,
1986);
6. Public expenditures are divided into two categories:
a) Social programs expenses including social security, income transfers, health,
education and leisure;
b) Public investment in infrastructure, capital goods, flooring, lighting and urban
design.
The influence diagram shown in Figure 3 presents the causal interaction between
the two regions (A and B) competing for attracting private investment into their
territory. The arrows indicate the causal relationships between variables in the model,
and the positive sign in the arrowhead indicates a direct relationship, while the negative
sign an inverse relationship. The arrows marked by two parallel bars have gaps (delays)
between the effects of change in a variable on the other variable.
Vv
.
0 ew \
Production A
Infraestructuré
& Unemployment
lice Rate A,
enetits A.
oA, bs
1 levestron Foe
Investment Aw
‘iB
Uniemployent
. — Rate B
. Te
une '
ws .
Tre Renonee B
4g Taken
Figure 3 - Diagram of the systemic interaction between two regions.
Source: Authors.
The logic contained in the diagram is a combination of archetypes analyzed
previously applied to the problem of tax war. Suppose there are only two regions and
that the total investment flow is determined exogenously, where (g;) is the proportion of
investment for the region A and (1—g,) is the portion allocated to the B region.
Assuming that, due to better allocation of infrastructure and logistics, the region
receives a greater proportion of private investment to the region B (steps I and II), this
leads to higher production of A (step III) (WILSON, Wildasin, 2004 .) The increased
production of A increases tax revenues in the region, increasing public spending on
infrastructure and social services (steps IV and V). The level of the infrastructure
increases after a delay (step VI) and consequently the production increase too, which
results in the first cycle of reinforcement (R1). With a better infrastructure the
attractiveness of capital also increases, resulting in a higher proportion of investment for
A (step VID.
The model proposed here is a zero sum so that increasing the proportion of
investment for the same reduction means in the portion intended to B, region B in
engaging a reverse path to the analogous region A. Therefore, the reduction in
investment in B reduces the capital stock, which decreases the production and public
spending on infrastructure (steps 1-6). After some time loops are triggered
reinforcements (R3 and R4), since the lower level of infrastructure restricts production
and reduce the attractiveness of investments. Thus, the region gets a head start (region
A) remains on a path of continuous distance to the less privileged state (region B), as
already provided the archetype for the successful succeed.
However, realizing his persistent disadvantage compared to their neighbors, the
policymaker decides to implement policies of tax incentives in the region B (step 7), in
order to attract more private investment into its territory and reduce unemployment. Tax
incentives, as the name suggests, stimulate investment in the territory for which it is
directed, increasing the proportion of investment in B, as well as the production level
after some time. The increased production tends to increase government revenue, but its
influence is balanced (step 8) for tax waivers. Assuming that the vector is superior to
the resignations of revenue resulting from increased production, we conclude that the
collection and spending of B decrease, leading, after a delay, poor infrastructure of the
state and the consequent reduction in the attractiveness of private investment ( Steps 4,
5, 6 and VII).
It should be added that, after the policymaker realize the strategy for the region
B, this does not hesitate to "fight back" in order that fails to capture the systemic effect
of the initiative to B - mentioned in the previous paragraph. So that both erupt into a
real "escalation" in search of a greater amount of private investment, as the archetype
admitted.
So here was shown a diagram of systemic tax competition, with the specification
of causal relationships in the form of stocks and flows. It now remains to introduce the
system of differential equations describing the system dynamics.
The production of each region is determined by a combination of factors capital
(K), manpower (L) and infrastructure (J) (KEEN; MARCHAND, 1997). The functional
form of production takes the specification of the cobb-douglas, with variable returns
(Wildasin, 2011, Wilson 1995). The equation which defines the production function is
shown below:
Ye = f (lr Ke be) = Le" Kee”) (1)
I= S (Boley a Di,) at (2)
K, = SFr, — Dep) at (3)
where /, = infrastructure; K, = capital stock; L, = manpower in each period.
The level of public infrastructure (J,) is determined by the integral of the
difference between public investment lag (Ip,_,) multiplied by a parameter (Bo) and
the depreciation of the infrastructure (D;,) - see equation (2). Since the capital stock
(K;) is given by the integral of the difference between the flow of private investment
(Fx,) and depreciation of capital (Dx,), according to equation (3). The variable
manpower (L,) in turn, is determined exogenously to the model, representing the
population of the region (ZODROW; MIESZKOWSKI, 1986).
The only source of government revenue is its own tax collection (Ar,), which, as
shown in (4), results from a combination of regional output (Y;) times the tax rate (Cr)
of the region minus the tax waivers (Rr,). The amount of tax waiver (Rr,) for the
current period is equal to the tax incentives lagged (If,_,), which were granted in the
previous period, as in (5).
Ar, = A — Rr )Cr¥e (4)
Rr, = Trp (5)
Given that public spending must respect the limits of its budget revenues and
public expenditures are divided into three categories:
Ar, = Gp, = Gs, + Ip, (6)
Where Gp, = public spending; Gs, = social spending and social services, and Ip, =
public investment.
Additionally, it is assumed that the proportion of spending allocated to public
investment (/p,), and social spending (Gs,) can be changed depending on the interest of
the ruler. However, due to the limited room for maneuver, a proportionate increase in
investment spending implies a reduction in social spending and vice versa. The
proportional distribution is expressed by following equations:
Ip = PoGp (8)
Ss = Gs = (1 — po)Gp, in wich 0 < py <1 (9)
Where pp = proportion of spending on public investment, and (1 — po) = the proportion
of social spending.
For the parameter (pg) is assigned the value of 0.22. For Rock and Giuberti
(2007), capital spending represents on average 22% of state budget expenditures. Thus,
other government spending, in this model, goes to social spending.
The unemployment rate (u,) obtained from equation (10) follows the principle
proposed by Okun's law, where the unemployment rate is the sum of the natural
unemployment rate (Uy) and the output gap (Yy — Y;) multiplied by the coefficient
adjusting (9). When the product is far from the effective potential, the unemployment
rate is far from the natural rate. If the actual output is above potential (Y, > Yy), the
unemployment rate is lower than the natural (4; < fy), the opposite being true
(MANKIW, 2004). To estimate the values of the difference (Yy — Y;) of the product
will be used a statistical procedure called the Hodrick-Prescott (HP), a smoothing
method widely used in literature to extract the cyclical component of the series
(ASSISL, DIAS, 2004).
Ht =o (Yy — ¥;) + Uy, onde xo> 0 (10)
Where fy = natural rate of unemployment; Yy = potential output.
The role of tax incentives granted by the regional governments to attract new
investment is described by combining two other functions: IF THEN ELSE and lookup.
The first is to provide a logical proposition as described below:
IF THEN ELSE{ ty S Mp
(true) => (lp, = 0] (e))
(false) = Ur, = flookup, G2) |]
The function attempts to test if the unemployment rate (,) is less than or equal
to the level tolerated by the policymaker (yp). If the proposition is true, the amount of
tax incentives will be null (Jp = 0), i.e. the policymaker will not compromise future
income to attract capital and create new jobs. On the other hand, if the unemployment
rate is not at acceptable levels, the policymaker puts the current tax incentives to attract
private investment.
The amount of tax incentives, i.e. the percentage of future revenues that the
government resigns, is determined by the equation/, = flookup, (=), whose
A
relationship is represented by the lookup function described in Figure 4. It is assumed
that, the higher the relative endowment of the neighboring region capital @, the
A
greater the incentives granted by the policymaker to attract private investment into its
territory. Thus, as it reduces the attractiveness of a region in relation to investment, this
region can extend tax incentives grants.
However, each region can not commit more than 6% of its revenue in future
incentives.
If
es ey
Source own elaboration.
Note: imput = values of the explanatory variable, output = value that the function can assume.
10
Finally, it remains to explain how the flow of capital is directed to each
jurisdiction. The value of the ratio of private directed to the region (g) is determined by
equation (12).
a= 7 (55) +0-MG AD ~05) (12)
To better understand this relationship, first consider the situation in which
regions do not grant tax incentives (g,; = v(“# -). Thus g will be influenced only by
Taptla
the relative endowment of infrastructure in each region multiplied by a conversion
factor (y). For example, assuming that y = 1 if both regions present an equivalent level
of infrastructure, the proportion of investment in each region will be the same (g, = 0,5)
- each region receives 50% of total investments However, as the region gains
displaying infrastructure superior to BO
Ge grows and increases the flow of
investment directed to A. The —_— occurs in a similar manner. According to Wilson
and Wildasin (2004) firms are benefiting from public spending in infrastructure and
because of this, their investments are concentrated in regions with better allocation of
this factor.
When tax incentives are granted, this is a sign that at least one government is not
satisfied with the way in which private investment are shared - the value that g, takes.
As a result, the function is an additional component, namely, G == ) -05): The
FA
function indicates the net effect of this compensation policy incentives resulting from
low investment attractiveness policymaker that some judges may exist in its territory.
This effect is given by the difference in the proportion of incentives by A in relation to
all involved in the two regions. If both regions give the same amount of incentives, this
has no effect. However, if the region B dispend a greater volume of incentive
(ce a D< 05) the amount of investment for A, ceteris paribus, will decrease, and
Fy
ethierwise a similar manner.
For the model does not incur explosive trajectories and values incongruent (a
mathematical indeterminacy) was added to the equation (equation 12) the functions IF
THEN ELSE. So if there are tax incentives grants qa- GS we - 05) =0: Thus,
equation 12 can be rewritten as:
IF THENELSE[ py, + Ieg, = 0
(Verdadeiro) = [g.=y (- an ) ] (13)
Fats) => Ca =r(“*-)+ 0-0 a = 98)1 }
2.3 Handling the model
11
The calibration procedure allows to adjust the parameters so that the endogenous
variables of the model adhering to real economic series. The adjustment of errors
simulated in relation to the actual values of each variable was obtained using the
calibration tool of the software Vensim, version 5.7.
The simulation model of tax competition in Brazil was calibrated considering
four sets of real data, which are: the GDP of the Southeast, Northeast regions and tax
revenue in the Southeast and Northeast. It should be noted that different weights were
adopted for each series in order that the differences in magnitude does not come to favor
certain series at the expense of others. After testing different combinations of weighing,
we opted for the combination shown in Table 2, for being who approached
simultaneously the series for the Southeast and Northeast.
Table 2 — Series’ weighing of the calibrated simulation model
Series Weight’s parameter Value
Southeast GDP wl 0,20
Northeast GDP w2 0,40
Southeast Tax Revenue w3 0,15
Northeast Tax Revenue w4 0,25
Source: Developed by the authors.
2.4 Source and data processing
The series used in the modeling of tax competition in Brazil in the Gross
Domestic Product (GDP) in Southeast and Northeast and the projection of the resident
population in each region as a proxy for hand labor were extracted from the database of
the Institute for Research in Applied Economics (IPEADATA). The series of state tax
revenue from the two regions (deflated by the implicit deflator of GDP) were collected
from the National Treasury Secretariat (STN). Additionally, it was used as a proxy for
private investment the sum of energy consumption (in Gigawatts) in the industrial
sector of both regions, obtained from the site of the Central Bank of Brazil. According
to Casali, Silva Carvalho (2010) power consumption is highly correlated with private
investment, since the more intense the investment in machinery and equipment, the
greater the energy consumption by industries.
All data correspond to an annual series; the period begins in 1985 and extends
until 2005. To standardize the units of measurement used in the simulation all series
were standardized in terms of the proportion of each variable in relation to the total
GDP of the regions in 2008. Exceptionally series relating to manpower were normalized
in terms of the proportion of the total population in each territory also distributed in
2008. Thus, the values assumed by each variable have been set as shown in Table 3.
12
Table 3 — Normalized values of each variable in 2008
Series Normalized value
Southeast GDP 81,04
Northeast GDP 18,96
Southeast Tax Revenue 6,84
Northeast Tax Revenue 1,71
Southeast Labor 60,17
Northeast Labor 39,83
Private Investment of the two regions 15
“Source: Developed by the authors.
3. Results of the simulation
3.3.1 Adjustment of the model
The structural behavior of the model of tax competition has been defined
through a system of differential equations and causal relationships presented in the
previous section.
The calibrated parameter settings and their values are shown in Table 3.
The parameters q,w and z correspond respectively to the coefficients capital return,
manpower and infrastructure of the production function. As the production function is
specified the type cobb-douglas, the coefficients are the production elasticity for each
factor.
The value assumed by the conversion coefficient of public investment in
infrastructure (8) indicates that the investments are similar in both regions.
Table 4 — Calibrated parameters of the fiscal model
Parameter Value Description
qo 0,5298 Capital return coefficient on production in Southeast
CH 0,4910 Capital return coefficient on production in Northeast
Wo 0,2144 Labor return coefficient on production in Southeast
Wy 0,1750 Labor return coefficient on production in Northeast
Zo 0,4600 — Infraestructure return coefficient on production in Southeast
zy 0,4600 —Infraestructure return coefficient on production in Northeast
Bo 0,9350 Transformation coefficient of public investment in infrastructure
in Southeast
By 0,9399 Transformation coefficient of public investment in infrastructure
in Northeast
Ho 0,1005 Natural unemployment rate in Southeast
HN, 0,0943 Natural unemployment rate in Northeast
cay 0,0034 Business cycles effects coefficient on Southeast unemployment
a 0,0076 Business cycles effects coefficient on Northeast unemployment
TB 0,0600 — State jurisdiction tax burden before 1988 constitution
TBA 0,0260 Tax burden change in Southeast after 1988 constitution
13
TBB 0.0290 tax burden change in Northeast after 1988 constitution
Gastos Pub. 0,8972 Initial value of public spending in Southeast
Delay A
Gastos Pub. 0,0100 Initial value of public spending in Northeast
Dealy B
Depreciagao 0,1000 — Infraestructure depreciation rate in Southeast
Infraestrutura A
Depreciagao 0,1000 — Infraestructure depreciation rate in Northeast
Infraestrutura B
Depreciagao 0,0743 Capital depreciation rate in Southeast
Capital A
Depreciagao 0,0656 Capital depreciation rate in Northeast
Capital B
Capital Delay A 60,000 _ Initial value of capital stock in Southeast
0
Capital Delay B 21,870 Initial value of capital stock in Northeast
4
Infraestrtura 6,5349 Initial value of infraestructure level in Southeast
Delay A
Infraestrutura 1,7783 Initial value of infraestructure level in Northeast
Delay B
Yy 0,8947 Adjustment coefficient of private investment in relation to
regional distribution of capital
Source: Search results.
Note: * The index 0 and 1 correspond to the Southeast and Northeast, respectively. A ** = Northeast. B =
*** Southeast. CT = **** tax burden.
We highlight the natural unemployment rate (uy) set the template for each
region. The calibrated values indicate that the Southeast region has a natural rate of
unemployment higher than the rate in the Northeast, although the values are not so
distinct. This difference in values is supported by the data presented by Pochmann
(1998) concerning the unemployment rate by geographic regions between 1989 and
1996, where the Southeast region had unemployment higher than the Northeast. The
parameter which represents the effect of economic cycles on the level of unemployment
was also lower in the Southeast region, indicating that this is ,ceteris paribus, less
susceptible to fluctuations in employment that the Northeast.
Regarding the tax burden was possible to detect a difference between the periods
before and after the enactment of the Federal Constitution (FC) in 1988, when states
acquire greater autonomy in the administration of taxes in their jurisdiction. The tax
burden (CT) model is adjusted by 6% until 1988, during which there is no distinction
between rates of states (Nascimento, 2008). Since 1989 both regions begin a gradual
increase in their tax burdens, but in distinct ways. The Southeast region stabilizes at a
value of the tax burden of 8.6% (CT + CTA), while the Northeast is now 8.9% (CT +
CTB). Although not distance themselves from this, it is noticed that there is a difference
between the prices charged by each region. This indicates that the states make use of
regions some autonomy in setting their tax jurisdiction. The fact that both regions
increase their level of tax burden, each in its own way, reflects the decentralized aspect
of the present Federal Constitution of 1988, which increased revenue sources in
detriment of States and the Union (VIOL, 1999).
14
The remaining parameters fitted represent the initial conditions of the model
relative to the stock of capital, level of public spending on infrastructure and investment
in each region, where the Southeast had higher values as the Northeast region. The
depreciation rates of capital and infrastructure in each region represent the only means
of escape (exit mechanism) resource level of these variables. As a result, it is possible
that the values of the depreciation rates are incorporated in other factors that affect the
output of those resources (such as capital mobility between regions and countries).
Finally, the parameter (y) measures the effect of different spatial distribution of
infrastructure on private investment decision while (1 — y) correspond the proportion to
the effect of regional tax incentives on private investment decisions. Thus, we conclude
that, according to the calibrated model, the influence proportion of infrastructure and tax
incentives in the private investment decisions in each region are respectively 89.5% and
10.5%.
Additionally, the parameter (up), which represents the level of unemployment
tolerated by the policymaker, was defined based on an estimate of the NAIRU for Brazil
developed by Silva (2008). According to him the true value of the Brazilian NAIRU is
in the range between 7.4% and 8.5%. For this reason, in this study was found a range
mean of up = 0,08.
Figures 5 and 6 shows the behavior of GDP and tax revenues series per region,
enabling the adhesion of the calibrated model to real series. The calibration have been
applied to four sets of real and potential fluctuations what allow the model capture
better the behavior of certain series, as seems to occur with GDP in the Northeast.
Despite any disturbances, the model has an acceptable visually degree of adhesion, so
that the calibrated model can be considered structurally representative.
GDP of the Southeast and Northeast ‘Tax Revenue of the Southeast and Northeast
100 3
5 6
50 = a 2:
2s 2
0 o
19851988 199119931997 2000 2003-2006 1985 1987 1989 1991 1993 1995 1997 1999 2001 2003 2005 2007
Time (Year) Time (Year)
Figure 5 — GDP’s and tax revenues real and modeled in the Southeast and Northeast,
1985-2008.
Source: Search results.
3.3.3 Analysis of the variables
The analysis period (1985-2005) is marked by deep political and economic
transformations in the national context. During the second half of the 1980s Brazil
experienced a process of democratization accompanied by failed attempts to stabilize
the price level, which came to be controlled only in mid-1994 with the Real Plan
(GIAMBIAG et al., 2005). The economic and political instability, together with the
growing pressure for fiscal adjustment and deficit reduction, not created a favorable
15
environment for investment in infrastructure. As shown in Figure 8 modeled levels of
infrastructure in the regions remain on a downward trend until 1995. Since 1996 the
regions expand the stock of infrastructure, with stronger growth in the Southeast. The
southeast had an average growth of 1.69% while the Northeast was 0.96% annually.
The Northeast region has less infrastructure that the Southeast region, so this
region has more attractiveness for new investments. Thus, in order to overcome this
structural disadvantage, the Northeast is to grant tax incentives to attract such
investment. As shown in Figure 9, the incentives are to be granted from 1989. Although
the structural disadvantage is attributed solely to the Northeast, both regions are to be
granted tax incentives. The explication are that the Southeast do not want to lose
potential investment installed in your jurisdiction. As a result, the Northeast region need
an additional effort (in terms of tax waiver) to become more attractive.
Infraestructure Level of the Southeast and Northeast
0
1985 1988 1991-1994 1997-2000 -—-2003-~2006
Time (Year)
Infraestructure of the Southeast: Modeled. —A—4A—A—_4A—_& 4&4
Infraestructure of the Northeast: Modeled —-8——8——8—8— 8-8-8 —
Figure 8 — Modeled level of Infraestructure in the Southeast and Northeast, 1985-2008.
Source: Developed by the authors.
Fiscal Benefits Granted by Southeast and Northeast
0.08
0.06
1985 1988 1991 19941997 2000-2003» -2006
Time (Year)
Fiscal Benefits Granted by Southeast: Modeled —A—A—A—A—_A—__ + —_*&
Fiscal Benefits Granted by Northeast: Modeled —-2—-8—8—8—8—\_8-—
Figure 9 — Fiscal benefits granted by the Southeast and Northeast, 1985-2008.
Source: Search results.
16
Note, however, that the beginning of the tax incentive awards coincides with the
expansion of the tax burden of the two regions - from 1989. That is, to circumvent the
COMMITMENT revenue from tax waivers, the government increased tax rates imposed
on society. It should be noted that tax competition encourage business groups awarded
tax incentives and burdens the local productive sectors already installed. This evil
character is more pronounced in the Northeast, since the increased tax burden and the
amount of incentives is higher.
Because of the incentives, the Northeast region extends (from 1992) the volume
of investments in its territory, as follows in Figure 10. It is also reduced, even so
significant bit, the distance of the volume of interest of each region, as shown in Figure
11. Thus, it is possible to infer that tax competition has helped to reduce disparities in
the spatial distribution of private investment, but its effects cannot systematically affect
the distribution, since the growth of the stock of infrastructure was impaired by tax
waivers and, consequently, reduced the attraction of private capital.
Investment Made in the Southeast and Northeast
12
9
6
Se O GEO ae oe coneeenanal
0
1985 198819911994 1997 2000-2003) 2006
Time (Year)
Investment Flow Directed to Southeast: Modeled &—A—A—A kA Ah
Investment Flow Directed to Northeast: Modeled. —8—8—8—8—8—8—_8—8
Figure 10 — Modeled investment flow for the Southeast and Northeast, 1985-2008.
Source: Search results.
Capital Endowment in the Southeast and Northeast
100
75
50
os TTT
0
1985 1988 1991 1994 1997 2000 2003 2006
Time (Year)
Capital Endowment in the Southeast: Modeled. —&’—&—A—& kA
Capital Endowment in the Norhteast: Modeled. —_3—s8—8—s—8—s—
Figure 11 — Modeled capital stock in the Southeast and Northeast, 1985-2008.
Source: Search results.
17
Despite the amount of investment attracted to the Northeast, the trend of tax
revenues remain as seen in Figure 12. The trajectory of unemployment patterned for the
regions are shown in Figure 13. It’s possible to check that the Southeast region as well
as presenting a higher level of unemployment in almost all periods, also shows a higher
rate of oscillation. This is because the region has major disturbances of GDP around its
potential value, i.e. the presence of the cyclical component. On the other hand, the GDP
of Northeast has more moderate behavior, allowing minor fluctuations in
unemployment around its natural value.
Tax Revenue of the Southeast and Northeast
ote HTH TTT TTL
1985 1987 1989 1991 1993 1995 1997 1999 2001 2003 2005 2007
Time (Year)
Tax Revenue of the Southeast: Modeled. —&—#&—A—A—& AA
Tax Revenue of the Northeast: Modeled ——8—s8—8—8—8—s—8—
Figure 12 — Modeled tax revenues of the Southeast and Northeast, 1985-2008.
Source: Search results.
Unemployment Trajectory in the Southeast and Northeast
0.15
0.1125 a AN
0.075
0.0375
0
1985 1988 1991-1994. 199720002003 2006
Time (Year)
Unemployment Rate in the Southeast: Modeled. <A —A—A—A AA
Unemployment Rate in the Northeast: Modeled. —s—8—8—8—8—8—
Figure 13 — Evolution of the modeled unemployment rate in the Southeast and
Northeast, 1985-2008.
Source: Search results.
3.3.4 Scenarios analysis
The purpose of this section is to evaluate the behavior of the main series
modeled upon the occurrence of certain transformations or changes of initial conditions,
18
which are classified as scenarios. The scenarios simulated in the model of tax
competition between Southeast and Northeast regions are described in Table 3. Scenario
1 shows the system behavior in the absence of tax incentives. The second scenario is a
standardization of state competence rates, ie. to demonstrate the model behavior in
absence distinctions between aliquots practiced by the regions. This standardization
would be similar to the tax reform proposal contained in the constitutional amendment
No. 42/2003 discussed by Paes and Smith (2004). According to this proposal, there
would be the unification of the laws of GST and replacing it with a value added tax in
order to extinguish the cascading effect of taxes and curb tax competition.
Box 3 — Simulated scenarios in the fiscal competition model
Scenario Scenarios description
Scenario | Prohibition of the granting tax benefits.
Scenario 2 Standardization of the tax burdens.
Scenario 3 Standardization of the tax burdens combined with the end of tax benefits.
Scenario 4 Redirection of the tax benefits for public investment.
Scenario 5 Redirection of the tax benefits for public investment combined with the
standardization of the tax burdens.
Source: Elaborated by the authors.
However, the mere standardization of tax rates does not address all the concerns
contained in the proposed tax reform. However, this standardization represents the
taxable percentage in the aggregate, not deducting the effect of waivers - granted for
specific sectors. It is therefore necessary to keep some room for some special treatment
- incentives to a portion of new investors. Accordingly, Scenario 3 is trying to meet
these goals by combining the standardization of tax rates with the ban on incentives.
The last two scenarios attempt to evaluate the hypothesis that regional
governments, instead of granting tax incentives for new businesses, invest the
equivalent sum in infrastructure in order to raise the regional attractiveness. The results
of the simulations with different scenarios are presented in Figures 14, 15,16, 17 and 18.
You can see that in all scenarios proposed in the Southeast results most
favorable than those observed in practice, while the Northeast got different results. In
the absence of specific policies to attract private investment, the Northeast cannot
mitigate this regional bias in the allocation of this resource, since the Southeast is
naturally more attractive to investment (scenario 1).
On account of lower investment flows directed to the Northeast, the capital stock
of the region has a lower growth trajectory. The trajectory of the capital stock has a
negative effect on GDP in the Northeast. Regarding tax collection and infrastructure
level, there is not a significant change in behavior. This is due to the absence of tax
waivers. Because they are not granted tax incentives, reduced tax base (decrease of
GDP) is offset by an increase in the percentage of capitation tax (no waivers). Thus,
both the inflow and the infrastructure exhibit paths similar to those observed in practice.
19
The Southeast region has a crescent private investment flow, and therefore
would not need to use their tax incentives. The only motivation in the Southeast to grant
tax incentives is the Northeast, that have started this practice, i.e. the Southeast reacts to
"attack" initiated by the Northeast in order to avoid losses of private capital. Thus, under
the conditions of scenario 1, the Southeast expands investments, amount of capital,
GDP, tax revenue and infrastructure, which tends to attract more private investment
(positive spiral).
The standardization of state competence rates (Scenario 2) also restricts the
room for maneuver in the Northeast. With the homogenization of tax burdens, the
Northeast region reduces the amount raised and therefore the amount of investment in
infrastructure. In return, the Southeast region increased its revenue and may increase
investments in infrastructure. However, this standardization does not change the spatial
distribution of the invested amount. For this reason, the result of scenario 2 does not
produce very significant changes in relation to GDP, and we found a reduction of GDP
Northeast and expansion of GDP in the Southeast.
In the third scenario the effect of the ban on tax incentives earn an additional: the
standardization of tax burdens. In the case of the Southeast, the tax burdens further
extends the storage, dispensing more resources for infrastructure investment (more
attractive), without the need for tax waiver. The Northeast, on the other hand, gets the
worst results in terms of flow of private investment, GDP and revenues. This is because,
besides the absence of policies to attract investment, reduced revenue limits the volume
of public investment in infrastructure, making the Northeast even less attractive for
investment (scenario 3). Thus, the GDP of Northeast also stays any levels below those
in other scenarios.
In the event that the amount of tax incentives is transferred to investment in
infrastructure (Scenario 4), the results for the Northeast region are better than the
scenarios that prohibit the granting of incentives. However, in relation to the flow of
investment these results are still below the achieved levels in the region when practiced
tax competition. The fact that incentives are not directly grant to investors make the
Southeast the preferred location for making investments. So even though the Northeast
can realize significant increases in the level of infrastructure, its disadvantage in
attracting investment persists (success to succeed).
On the other hand, the gains made in infrastructure - Scenario 4 - increased
values achieved by GDP, both the Northeast and Southeast, indicating that policies
directed to the development of infrastructure tend to boost the economy's productive
capacity. With this argument, Malliagros and Ferreira (1998), evaluated the effect of
investments in infrastructure on TFP Brazilian find that there is a strong relationship
between infrastructure and output in the long time. According to the authors, with the
financial deterioration of the state, increasing indebtedness and acceleration of inflation,
the state and public investment have a marked reduction, which cause a reduction in the
GDP growth rate. Thus, the gap in the infrastructure levels has a very significant impact
of the GDP.
20
Finally, the redirection of tax incentives in conjunction with the standardization
of tax rates (scenario 5) results in a condition similar to scenario 4 in terms of attracting
investment. But the above scenario reinforces the results of scenario 4, intensifying the
expansion of GDP and stock of infrastructure in the Southeast, and mitigating its effect
on the Northeast.
Although the scope of this study are not a long-term analysis, we can infer that
with a constantly increase of infrastructure, the capital attractiveness of both regions
also increases. That is not a sufficient condition for delay reduction compared to the
Northeast. The sub national financial condition enrages the composition of public
spending. Since there is already a push for the performance of governments in areas
such as health and education, it becomes unviable to earmark a significant amount of
resources for development of infrastructure. In addition there are also costs involved in
coordinating these policies at regional level, with conflicting interests and opposition
politicians. These factors fall into this problem within the theme of collective action,
discussed in the previous chapter.
In the national point of view tax competition only results in loss of tax revenue
and distortions in the efficient allocation of resources (investment decisions are not
guided by differences in competitiveness), however in the regional point of view for
marginalized regions industrial policy are an alternative for private investment. In the
absence of a central action and planned reduction of regional disparities, the tax waiver
is revealed as a practical tool with immediate result. It should be noted that according to
the systemic model, the results obtained with the tax competition is not even optimal
condition for the less favored resource (Northeast). Since, upon the incursion of specific
investments in infrastructure (in the same amount of tax waivers), the behavior of GDP,
infrastructure and revenue would be higher in this region.
Modeled Investment Flow in the Southeast Modeled Investment Flow in the Northeast
‘sated in ed
a) b)
Invesament: Cliatad
Investment: Soi $
Figure 14 — Modeled investment flow for the Southeast and Northeast in the different
scenarios, 1985-2005.
Source: Elaborated by the authors.
21
Modeled Endowment of Capital in the Southeast Modeled Endowment of Capital in the Northeast
tos 8
905 ws
” u
es ns
so “
19831988 1991 994 1997 2000 2003 2006 19851988 1981 T998 1997 3000 — 2007 2006
Tine (Year) Tine (ear)
a) b)
Capital: Caine
Capital: Scenario 1
Capital: Scenario 2
Capital: Secaro 3 ~
Capital: Scenario 4
Capital: Scenario $
Figure 15 — Modeled capital stock in the Southeast and Northeast in the different
scenarios, 1985-2005.
Source: Elaborated by the authors.
Modeled GDP in the Southeast Modeled GDP in the Northeast
80.5 183
Raz 16.22 Lr
63.75 14s
55.37 12.07
4 10
198519881991 19941997 2000 2003 2006 198519881991 1994 1997 20002003 2006
Tine (Year) Time (Year)
a) b)
GDP; Calibrated
GDP: Scenario |
GDP: Scenario 2
GDP: Scenario 3 S
GDP: Scenario 4
GDP: Scenario
Figure 16 — Modeled GDP of the Southeast and Northeast in the different scenarios,
1985-2005.
Source: Elaborated by the authors.
Modeled Tax Revenue in the Southeast Modeled Tax Revenue in the Northeast
7 135
5.95 131s
49 1.08
3.85 os4s
28 061
loss 19881991 -1994_—:1997 2000-2003 2006 19851988 19911994 1997 200020032006
‘Time (Year) ‘Time (Year)
a) b)
Tae Revue See?
Figure 17 — Modeled tax revenue of the Southeast and Northeast in the different
scenarios, 1985-2005.
Source: Elaborated by the authors.
22
Modeled Level of Infraestructure in the Southeast Modeled Level of Infraestructure in the Northeast
978 Z 2525
85 195
725 1378
loss 1988 1991-1994 1997 2000 2003 2006 19851988 19911994 1997 30002003 2006
‘Time (Year) ‘Time (Year)
a) b)
Infaestructure: Calibrated
Infaestuctue: Semario 1
Infzstructure: Seanaro 2
Infaestructure: Senaro 3
Infaestructure: Senario 4
Infaestructue: Seonario 5
Figure 18 — Modeled level of Infraestructure in the Southeast and Northeast in the
different scenarios, 1985-2005.
Source: Elaborated by the authors.
5. Conclusion
The use of tax competition has become customary in the Brazilian federation.
Through the commitment of measures of fiscal and financial nature, sub national
governments seek to reverse the centripetal tendency of Brazilian politics, which
channels the flow of resources to the central regions. However, the lack of regulation
means able to coordinate its implementation provides status of conflict with this
phenomenon. In the present study seeking to analyze the interaction between systemic
and institutional sources of tax competition in Brazil, aiming to identify the elements
that perpetuate this practice and its effects on the economy.
The used system model indicates that in absence of tax competition, the
peripheral regions (Northeast) receive a lower volume of private investment, resulting in
a lower level of GDP. However, the scenario in which the volume of tax incentives
(which would be potentially granted) are channeled in the form of public investment in
infrastructure, it seen a trend of higher GDP growth of both regions. It should be noted
that among the scenarios that are not granted tax incentives, this is the condition
because the Northeast region has a higher flow of investment. However, its
implementation is barred by the inability of regional governments to promote systematic
investments in infrastructure, due to fiscal rigidity and coordination costs.
For this reason, it is evident that the sub national level (decentralized) is the only
alternative industrial policy for less grant tax incentives. However, this type of policy is
palliative and temporary and does not constitute, per se, a sustainable policy to fix the
route of economic contraction. It remains the central government to put policies in place
to increase the attractiveness of these regions for private investment. Otherwise, these
regions do not have incentives to abdicate tax competition.
The proposed and analyzed model in this study captures regional disputes for
new investments, but cannot measure the effect of mobility of investments already
installed or to influence investment decisions across countries, i.e. that tax competition
23
is important for attracting foreign investment. Another aspect is that the analyzed model
captures the aggregate behavior of tax competition between two regions. However, the
main actors of this phenomenon are the states and municipalities, so it is important to
analyze the nuances of the competition between various jurisdictions, which may have
similarities (social problems, fiscal imbalance, lack of infrastructure, etc). Additionally,
the model does not include mobility and heterogeneity of technical manpower (different
levels of skilled labor). Further work may consider these presumptions and incorporate
the above aspects.
References
ARBIX, G. Guerra Fiscal e Competigao intermunicipal por novos investimentos no
setor automobilistico brasileiro. Dados v.43 n.1. Rio de Janeiro, 2000.
ASSIS, Luciana C. de; DIAS, Joilson. Politica fiscal, nivel tecnolégico e crescimento
econdmico no Brasil: teoria e evidéncia empirica. Disponivel em:
<www.anpec.org.br/encontro2004/artigos/A04A050.pdf>. Acesso em: 20 Out. 2008
BCB — Banco Central do Brasil. Séries temporais. Disponivel em:
<http://www.beb.gov.br/>. Acesso em: 20 Jun. 2009.
BUENO, Newton Paulo. Introdugao 4 dinamica de sistemas. Vicosa: Ed. UFV, 2011.
. Institutions and natural resources sustainability in poor socio-ecological
systems: closing some dynamical loops. The Journal of Interdisciplinary Economics.
Vol. 21, 2009.
CASALI, Giovana F. Rossi. SILVA, Orlando Monteiro da. CARVALHO, Fatima M. A.
Sistema regional de inovacao: estudo das regides brasileiras. Revista de Economia
Contemporanea, Rio de Janeiro, vol.14 no.3 Sept./Dec. 2010
DULCL, O. S. Guerra fiscal, desenvolvimento desigual e relagées federativas no Brasil.
Revista de Sociologia e Politica. N.18 Curitiba jun. 2002.
FERREIRA, Pedro Cavalcanti; MALLIAGROS, Thomas Georges. Impactos produtivos
da infra-estrutura no Brasil — 1950/95. Pesquisa e Planejamento Econdémico. Rio de
Janeiro, v 28, nn. 2, =p 315-338, 1998. Disponivel em: <
http://ppe.ipea.gov.br/index.php/ppe/article/viewFile/708/648>. Acesso em: 09 fev.
2012.
FORRESTER, Jay W. Industrial Dynamics. Waltham, MA: Pegasus Communications
, 1961, 464pp.
GIAMBIAG, Fabio... [et al.]. Economia brasileira contemporanea. Rio de Janeiro:
Elsevier, 2005, p. 425.
GILBERT, N. and TROITZSCH, K. G. Simulation for the Social Scientist. Second
edition. McGraw-Hill, Open University Press, 2006.
IPEADATA. Dados macroeconémicos e regionais. Disponivel em: <
http://www. ipeadata.gov.br/>. Acesso em: 28 mai. 2011.
24
JERVIS, R. System Effects: Complexity in political and social life. Princeton, NJ:
Princeton University Press, 1997.
KEEN, M.; MARCHAND, M. Fiscal competition and the pattern of public spending.
Journal of Public Economics 66, 33-53, 1997.
KIM, D. H. Systems thinking tools: a user’s reference guide. Waltham: Pegasus, 2000.
p.3-21.
LUPORINI, Viviane; ALVES, Joana. Investimento privado: uma andalise empirica para
o Brasil. Economia e Sociedade, Campinas, v. 19, n. 3 (40), p. 449-475, dez. 2010.
MANKIW, N. Gregory. Macroeconomia. 5° Edicao . Rio de Janeiro: LTC, 2004. 379p.
MINTZ, Jack; TULKENS, Henry. Optimality Properties of Alternatives Systems of
Taxataion of Foreign Capital Income. Journal of Public Economics 60, 373-399, 1996.
NASCIMENTO, S. P. Guerra fiscal: uma avaliagéo comparativa entre alguns estados
participantes. Economia Aplicada. vol.12, n.4, pp. 677-706. 2008.
Ostrom, E. Collective action and the evolution of social norms. Journal of Economic
Perspective, vol. 14 (3), 2000.
PAES, Nelson Leitaéo e SIQUEIRA, Marcelo Lettieri. Reforma Tributaria no Brasil:
uma_andlise dindmica dos efeitos macroecondmicos e dos impactos sobre a
distribuigao de renda e 0 bem-estar. Brasilia: ESAF. Monografia premiada em
2° lugar no IX Prémio Tesouro Nacional — 2004, Topicos Especiais de
Finangas Publicas, Recife (PE), 2004. P. 80.
POCHMANN, Marcio. Velhos e novos problemas do mercado de trabalho no Brasil.
Indicadores Econémicos FEE, Vol. 26, No 2, 1998.
ROCHA, Fabiana. GIUBERTI, Ana Carolina. Composicgéo do gasto publico e
crescimento econdémico: uma avaliagéo macroeconémica da qualidade dos gastos dos
Estados brasileiros. Econonmia Aplicada, Sao Paulo, v. 11, n. 4, p. 463-485, outubro-
dezembro, 2007.
SILVA, Tito Nicias Teixeira da. Searching for the Natural Rate of Unemployment in a
Large Relative Price Shocks Economy: the Brazilian Case. Working paper series, No.
163, 2008, pp. 1-51.
SILVA, F. A. R.; OLIVEIRA, F. A. Descentralizacao e federalismo fiscal no Brasil:
desafios da reforma tributaria. Rio de Janeiro: FGV: Konrad-Adenauer-Stiftung,
2007.
STERMAN, John. Businness dynamics: systems thinking and modeling for a complex
world. Boston: Irwin McGraw-Hill, 2000.
STN — Secretaria do Tesouro Nacional. Execugéo Orcamentarioa por estado.
Disponivel em: <http://www.tesouro.fazenda.gov.br/>. Acesso em: 28 mai. 2011.
WILDASIN, David E. Fiscal competition for imperfectly-mobile labor and captal: a
comparative dynamics analysis. Journal of Public Economics 95, 1312-1321, 2011.
WILSON, John Douglas. WILDASIN, David E. Capital tax competition: bane or boon.
Journal of Public Economics 88, 1065-1091, 2004.
25
. Mobile Labor, Multiple Tax Instruments, and Tax Competition. Journal of
Urban Economics 38, 333-356, 1995.
VENSIM® — VENTANA SYSTEMS. Vensim User’s Guide: version 5. Inc.,
Massachucetts, 2007. Disponivel em: < http://www.vensim.com/causal.html>. Acesso
em: 18 fev. 2011.
VIOL, A. L. O fenédmeno da competic¢ao tributaria: aspectos tedricos e uma analise
do caso brasileiro. Brasilia: ESAF, 1999. 79 p. Monografia vencedora em 1° Lugar no
IV Prémio de Monografia - Tesouro Nacional, Topicos Especiais de Finangas Publicas,
Brasilia (DF), 1999.
ZODROW, G.; MIESZKOWSKI, P. (1986). Pigou, Tibout, Property taxation, and the
underprovision of local public goods. Journal of Urban Economics 19, 356-370, 1986.
26
