the effects of brazil's taxation and social spending on the distribution of household income

http://pfr.sagepub.com/Public Finance Review

http://pfr.sagepub.com/content/42/3/346The online version of this article can be found at:

DOI: 10.1177/1091142113501714

2013 2014 42: 346 originally published online 12 SeptemberPublic Finance Review

Sean Higgins and Claudiney PereiraDistribution of Household Income

The Effects of Brazil's Taxation and Social Spending on the

Published by:

http://www.sagepublications.com

can be found at:Public Finance ReviewAdditional services and information for

http://pfr.sagepub.com/cgi/alertsEmail Alerts:

http://pfr.sagepub.com/subscriptionsSubscriptions:

http://www.sagepub.com/journalsReprints.navReprints:

http://www.sagepub.com/journalsPermissions.navPermissions:

http://pfr.sagepub.com/content/42/3/346.refs.htmlCitations:

What is This?

- Sep 12, 2013OnlineFirst Version of Record

- Apr 3, 2014Version of Record >>

at Dicle Ãœniversitesi on November 11, 2014pfr.sagepub.comDownloaded from at Dicle Ãœniversitesi on November 11, 2014pfr.sagepub.comDownloaded from

http://pfr.sagepub.com/

http://pfr.sagepub.com/content/42/3/346

http://www.sagepublications.com

http://pfr.sagepub.com/cgi/alerts

http://pfr.sagepub.com/subscriptions

http://www.sagepub.com/journalsReprints.nav

http://www.sagepub.com/journalsPermissions.nav

http://pfr.sagepub.com/content/42/3/346.refs.html

http://pfr.sagepub.com/content/42/3/346.full.pdf

http://pfr.sagepub.com/content/early/2013/09/11/1091142113501714.full.pdf

http://online.sagepub.com/site/sphelp/vorhelp.xhtml



Article

The Effects ofBrazil’s Taxationand Social Spendingon the Distributionof HouseholdIncome

Sean Higgins1 and Claudiney Pereira1

AbstractRelative to other countries in Latin America, Brazil has high rates oftaxation and large social spending. We estimate the redistributive effect offiscal policy on income distribution and poverty in Brazil using householdsurvey data that contain detailed information about many labor andnonlabor income sources, direct taxes paid, contributions to the pensionsystem, transfers received, use of public education and health services, andconsumption. On the spending side, we find that although Brazil has somewell-targeted antipoverty programs, these transfers have relatively lowper capita amounts and a large portion of direct transfer beneficiaries arenonpoor. As a result, inequality and poverty reduction are low relative toBrazil’s spending. On the tax side, indirect taxes paid by the poor oftensurpass the direct transfer and indirect subsidy benefits they receive.

1 Tulane University, New Orleans, LA, USA

Corresponding Author:

Claudiney Pereira, Tulane University, 206 Tilton Hall, 6823 St. Charles Avenue, New Orleans,

LA 70118, USA.

Email: [email protected]

Public Finance Review2014, Vol. 42(3) 346-367

ª The Author(s) 2013Reprints and permission:

sagepub.com/journalsPermissions.navDOI: 10.1177/1091142113501714

pfr.sagepub.com

at Dicle Ãœniversitesi on November 11, 2014pfr.sagepub.comDownloaded from

http://www.sagepub.com/journalsPermissions.nav

http://pfr.sagepub.com


Keywordsfiscal policy, poverty, inequality, Brazil

Historically, Brazil has had one of the highest levels of inequality in the

world; in 1989, its Gini coefficient of 0.63 was second only to Sierra Leone

(Ferreira, Leite, and Litchfield 2008). Over the last decade, however,

inequality has been falling in Brazil, as in other countries in Latin America

(Lustig, Lopez-Calva, and Ortiz-Juarez 2013). Indeed, inequality has fallen

in Brazil in every year since 2001.1 The recent decline is largely due to

increased public cash transfers (Barros et al. 2010) and a more equal distri-

bution of educational attainment resulting from expanded access to educa-

tion in the 1990s (Gasparini and Lustig 2011). Social spending has become

both larger and more progressive (Silveira et al. 2011). Poverty has

decreased in every year since 2003 despite the recent recession.2 Brazil’s

conditional cash transfer program Bolsa Famılia is very effective at reduc-

ing poverty (Soares 2012), especially in rural areas (Higgins 2012).

We estimate the redistributive effect of fiscal policy in Brazil using the

Pesquisa de Orcamentos Familiares (POF), 2008–2009. In particular, we

estimate the effects of taxation (direct and indirect) as well as cash transfers,

indirect subsidies, and in-kind benefits on income distribution and poverty.

The rich detail of our data set allows us to single out the effects of each

direct tax and transfer without needing to simulate most taxes or benefits.

This has the advantage that unlike incidence studies based on microsimula-

tion models, our study is based on what individuals actually pay and receive

(assuming they report correctly), rather than what tax and program rules

dictate they should pay. There are many reasons that the incidence of the

fiscal system could be different in practice than in theory, such as evasion,

exclusion, and leakages.3

Our contribution is to undertake a comprehensive incidence analysis for

Brazil, including indirect taxes, energy subsidies, and in-kind benefits from

public education and health, to assess the distributive impact of various fis-

cal interventions and the fiscal system as a whole. By using a consistent

methodology (see Lustig, Pessino, and Scott 2014), the results for Brazil are

comparable to those of other countries. In comparison to the other countries

included in this issue, Brazil has relatively high taxation and spending, but

low inequality and poverty reduction relative to its spending. Some pro-

grams, such as Bolsa Famılia and Benefıcio de Prestacao Continuada

(BPC), are well targeted, but they make up a small share of social spending.

Higgins and Pereira 347



Others, such as unemployment benefits and special circumstances pensions,

are large and progressive only in relative terms. While public health spend-

ing is progressive in absolute terms for each type of care, tertiary education

spending is almost neutral in relative terms, indicating that the better-off

receive most of the benefits.

Overall, direct taxes and transfers reduce the Gini by 6 percent, and in-

kind transfers are particularly equalizing: the reduction between the market

income and final income Ginis is 24 percent. Although Brazil’s market

income Gini is substantially higher (by at least 5 percentage points) than

that of any of the other countries included in this special issue, its final

income Gini is lower than Bolivia’s and Peru’s. Indirect taxes have a dele-

terious effect on post-fiscal income and often result in post-fiscal income

poverty being higher than market income poverty.

The article is organized as follows: the next section describes the social

spending and taxation systems in Brazil. The third section describes the data

used as well as the methodology, focusing on aspects of the methodology that

are unique to Brazil. The fourth section summarizes the main results of our

incidence analysis. Some conclusions are presented in the fifth section.

Social Spending and Taxation in Brazil

Social Spending

Social spending accounts for 16 percent of gross domestic product (GDP) in

Brazil when contributory pensions are not included. This figure includes

social assistance (direct transfers and other social assistance), health spend-

ing, and education spending at the federal, state, and municipal levels. If we

also include spending on contributory pension payments, social spending is

25 percent of GDP. Direct transfers include conditional cash transfer pro-

grams, noncontributory pensions, food transfers, unemployment benefits,

special circumstances pensions, and others. In-kind transfers are benefits

received from the universal free public education and health systems. The

main programs are described later, and their budget sizes are given in table 1.

Bolsa Famılia, Brazil’s flagship conditional cash transfer program,

transfers cash to eligible families in exchange for complying with certain

conditions. Eligible families are poor families with children less than eigh-

teen years of age or with pregnant women, and all extreme poor (the latter

group is regardless of having children). Eligibility is determined through

partially verified means testing; households with income below the cutoffs

are incorporated into the program. The conditions are prenatal and postnatal

348 Public Finance Review 42(3)



Table 1. Brazilian Social Spending, 2009.

Spending componentIncluded in

analysisBillions of

ReaisPercentage

of GDPNotes and

source

Direct cash and food transfersSpecial circumstances pensions Yes 72.6 2.3 hUnemployment benefits Yes 18.6 0.6 eBPC (noncontributorypensions)

Yes 16.9 0.5 a

Bolsa Famılia (CCT) Yes 12.5 0.4 aAssistance from PIS/PASEP Yes 7.3 0.2 eScholarships Yes 3.5 0.1 fOther elements of basic socialprotection

No 2.4 0.1 b, c

Food for workers program No 0.5 0.0 eBolsa Escola, Auxılio Gas, andother auxılios

Yes 0.4 0.0 b

Other food access programs No 0.4 0.0 c, gChild labor eradication Yes 0.3 0.0 bMilk transfer program Yes 0.2 0.0 gMinimum income programs Yes 0.1 0.0 dProfessional qualification grant No 0.1 0.0 eBasic food basket Yes 0.0 0.0 g

Social assistance (not direct transfers)Assistance to the elderly anddisabled

No 19.0 0.6 i

Community assistance No 18.1 0.6 iOther No 4.3 0.1 iAssistance to children andadolescents

No 2.7 0.1 i

EducationPrimary education Yes 75.1 2.4 iOther Yes 46.5 1.5 iTertiary education Yes 26.0 0.8 iSecondary education Yes 12.0 0.4 iEarly childhood education Yes 9.6 0.3 i

HealthInpatient care Yes 81.7 2.6 iOther Yes 41.6 1.3 iPrimary care Yes 33.6 1.1 iPreventative care Yes 9.1 0.3 iSocial spending analyzed(benchmark)

Yes 467.4 14.7

Total social spending(benchmark)

Part 515.1 16.2




care sessions for pregnant women, adherence to a calendar of vaccinations

for children ages zero to five, and a minimum level of school attendance for

children ages six to seventeen. There are no conditions for the ‘‘fixed ben-

efit’’ given to extremely poor households. There were 41.2 million individ-

uals living in beneficiary families in 2009, and the average benefit per

person living in a beneficiary household was US$0.35 per day adjusted for

purchasing power parity (PPP).4

BPC (Continued Payment Benefits) is a noncontributory pension pro-

gram that provides a monthly monetary transfer of one minimum salary

(465 Reais per month [US$8.83 PPP per day] in 2009) to elderly poor or

Table 1. (continued)

Spending componentIncluded in

analysisBillions of

ReaisPercentage

of GDPNotes and

source

Contributory pensionsFederal contributory pensions(INSS)

Yes 164.8 5.2 j

State contributory pensions Yes 56.1 1.8 iOther federal contributorypensions

Yes 53.7 1.7 c, i

Municipal contributorypensions

Yes 14.0 0.4 i

Social spending analyzed(sensitivity analysis)

Yes 756.0 23.7

Total social spending(sensitivity analysis)

Part 803.7 25.2 k

Source: Authors’ elaboration based on Pesquisa de Orcamentos Familiares, 2008�2009.Note: BPC ¼ Benefıcio de Prestacao Continuada; CCT ¼ conditional cash transfer; GDP ¼gross domestic product. All spending totals include spending at the federal, state, and municipallevels, unless otherwise specified. (a) Amount paid in transfers; see http://aplicacoes.mds.-gov.br/sagi/miv/miv.php. (b) Ministerio do Trabalho (2011). (c) Calculated as a residual by theauthors. (d) This is the total for Renda Cidada in Sao Paulo state, which is the largest minimumincome program. Secretaria do Desenvolvimento Social, Governo do Estado de Sao Paulo. (e)Ministerio do Trabalho (2011). (f) Portal da Transparencia, Controladora Geral da Uniao. (g)Programa de Aquisicao de Alimento, Companhia Nacional de Abastecimento (Ministerio daAgricultura, Pecuaria e Abastecimento). (h) This is the total for pensoes and outros benefıciosfrom Relatorio de Gestao do Instituto Nacional do Seguro Social (Ministerio da Previdencia eAssistencia Social). (i) Balanco do Setor Publico Nacional, Secretaria do Tesouro Nacional(Ministerio da Fazenda). (j) This is the total for aposentadorias and benefıcio mensal from Rela-torio de Gestao do Instituto Nacional do Seguro Social (Ministerio da Previdencia e AssistenciaSocial). (k) This number can be compared with Brazil’s total social pending as a percentage ofGDP according to the UN Economic Commission for Latin America and the Caribbean of 27percent.




incapacitated poor. Elderly means sixty-five years old and older, and inca-

pacitated is determined by doctors based on ability to work. In 2009, there

were 3.2 million beneficiaries,5 and the average benefit per person living in

a beneficiary household was US$2.18 PPP per day.

Unemployment insurance is funded by taxes on employers through the

Fundo de Amparo ao Trabalhador (Worker’s Assistance Fund). Eligibility

requirements include working continuously for at least six months prior to

the layoff and not receiving BPC. The benefit varies according to worker’s

salary, ranging from 1 to 1.9 minimum monthly salaries with a maximum of

five monthly payments. There were about 8 million beneficiaries in 2009

(Ministerio do Trabalho 2011), and the average benefit per person living

in a beneficiary household was US$0.74 PPP per day.

Special circumstances pensions are funded by the contributory pension

system, but they are considered noncontributory because they have low or

no requirements in terms of length of contribution period and are designed

to smooth the impact of idiosyncratic shocks or are means tested. They are

paid in the case of an accident at work, sickness, or related idiosyncratic

shock. In 2009, there were about 2.9 million beneficiaries (Instituto Nacional

de Seguro Social [INSS] 2010) and the average benefit per person was

US$5.22 PPP per day. Retirement and disability pensions, on the other hand,

are not considered part of social spending in the benchmark scenario (but are

treated as a government transfer in sensitivity analysis 1).6 The retirement age

in urban areas is sixty-five for men and sixty for women; in rural areas, it is

sixty and fifty-five, respectively. Alternatively, people under the retirement

age can still receive benefits, if they have contributed for a minimum of

thirty-five years for men and thirty years for women. In 2009, there were

7.6 million beneficiaries receiving benefits based on retirement age and 4.2

million receiving benefits based on contribution time (INSS 2010), and the

average benefit per person living in a beneficiary household was US$7.22

PPP per day. In the benchmark scenario, special circumstances pensions are

considered a government transfer, while contributory pensions are considered

part of market income. In sensitivity analysis 1, we consider contributory

pensions as a government transfer along with special circumstances pensions.

In sensitivity analysis 2, we consider both contributory and special circum-

stances pensions as part of market income.

Food transfer programs (Programa de Aquisicao de Alimentos [PAA])

provide food to low-income households. In the largest program, PAA Leite,

the government purchases milk from farmers and provides it for free to poor

households in the nine states of Brazil’s Northeast region and part of Minas

Gerais state. To be eligible, a household must have household per capita




income of less than half the monthly minimum salary and have at least one

of a child between two and seven years old, a pregnant or breast-feeding

mother, or an elderly person over age sixty. Eligible households receive one

or two liters of milk per day.

The main indirect subsidy in Brazil, the Social Tariff on Electric

Energy (TSEE), is a price subsidy on energy for low-income households

with total energy consumption below 220 kilowatt hours (kWh) per

month. In 2009, eligible households consuming less than 30 kWh per

month received a 65 percent discount, households consuming over 30 but

less than 100 kWh received a 40 percent discount, and households con-

suming between 100 kWh and the regional limit (at most 220 kWh)

received a 10 percent discount. All households consuming less than 80

kWh were eligible, while households consuming between 80 and 220 kWh

were required to have income below 120 Reais per month (US$2.30 PPP

per day) per capita to be eligible. The average benefit per person in a ben-

eficiary household was US$0.36 PPP per day.

Public education in Brazil is free at all education levels, including pre-

school and tertiary education. There is also free public day care provided

for poor families. The large majority of Brazilians attending school are

enrolled in the public system: 85 percent of elementary students, 86 percent

of secondary students, and 75 percent of postsecondary students. Public

health care is free for all types of care: instead of a national health insurance

system, as is common in many countries, Brazil has the Unified Health Sys-

tem (SUS) created by the 1988 Constitution, which guarantees free and

unlimited access to health care to every citizen at public facilities, and is

fully tax-financed.

The Brazilian Tax System

The Brazilian tax system is exceedingly complex, with more than eighty-

five taxes (Portal Tributario 2012). Total fiscal revenues at the federal,

state, and municipal levels were about 35 percent of GDP in 2009. Direct

taxes represent 45 percent of the taxes levied by the government and indi-

rect taxes represent 55 percent. Individuals are required to file personal

income tax returns, if taxable income goes above the exemption limit of

about 3.3 times the monthly minimum salary; their marginal rates range

from 15 to 27.5 percent. Because of the high exemption threshold and large

informal sector, less than 10 percent of the economically active population

pays personal income taxes (Immervoll et al. 2009). Corporate taxable




income is taxed at 25 percent. In addition, businesses must pay social con-

tribution taxes on profits (9 percent on net taxable income).

Many indirect taxes operate each with their own administering depart-

ment which may be at the federal, state, or municipal level. The most impor-

tant indirect tax is the Imposto sobre Circulacao de Mercadorias e Servicos

(ICMS), a state tax levied on the sale or physical movement of goods, freight,

transportation, communications services, and electricity. Intrastate transac-

tions are taxed at 18 percent on average (with rates determined at the state

level), interstate transactions at 7 percent or 12 percent (regulated by the

Brazilian Senate), and imports at a rate between 4 percent and 25 percent.

Communication services are taxed at a rate between 13 and 25 percent. ICMS

accounted for 21 percent of fiscal revenues in 2009. Other important indirect

taxes are the Contribuicao para o Financiamento da Seguridade Social

[COFINS (federal tax on goods and services to finance the social security

deficit)], Imposto Sobre Servicos [ISS (municipal tax on services)], Pro-

grama de Integracao Social [PIS (federal tax on goods and services to finance

social services for workers)], and Imposto sobre Productos Industrializados

[IPI (federal tax on industrial products)]. They correspond to 10.8, 4.1, 2.9,

and 2.8 percent of fiscal revenues, respectively (table 2).

The cascading effect, one of the Brazilian system’s major distortions,

derives from the fact that taxes levied at the federal, state, and municipal

levels compound on each other. This occurs because the taxes are applied

to the final sales price of the good (including taxes), not the pretax sales

price. About 18 percent of the collected revenue in 2003 was attributed

to compounded taxes resulting from the cascading effect (Siqueira,

Nogueira, and Souza 2010), and the overall cost of the distortions created

by it was about 2 percent of GDP (Amaral, Olenike, and Amaral 2007).

As we are analyzing the effects of fiscal policy on income inequality and

poverty, the distortions created are especially important due to their effects

on consumer purchasing power. Exemptions on consumption taxes are

almost nonexistent in Brazil (Corbacho, Cibils, and Lora 2013); hence, the

effective rates paid on basic food products in Brazil can be especially dele-

terious for the poor.

Data

The data on household incomes, taxes, and transfers come from the POF

(Family Expenditure Survey), 2008–2009. This survey has national cover-

age, sampling 56,091 households using a two-stage stratified sample

design, and is conducted approximately once every five years. It contains




Table 2. Brazilian Fiscal Revenue, 2009.

Included inanalysis

Billions ofReais

Percentage oftotal

Percentage ofGDP

TaxesFederal

Corporate income tax(IRPJ)

No 124.6 11.4 3.9

Tax on goods/services tofinance pensions(COFINS)

Yes 117.9 10.8 3.7

Individual income tax(IRPF)

Yes 67.1 6.1 2.1

Payroll tax collected fromemployers (FGTS)

Yes 54.8 5.0 1.7

Others No 46.9 4.3 1.5Contribution on net

profit (CSSL)No 44.2 4.0 1.4

Tax on industrializedproducts (IPI)

Yes 30.8 2.8 1.0

Tax to finance socialservices for workers(PIS)

Yes 31.8 2.9 1.0

Tax on financialtransactions (IOF)

No 19.2 1.8 0.6

Imported goods No 16.1 1.5 0.5Tax on technical services

(CIDE)No 4.8 0.4 0.2

Tax on rural properties(ITR)

Yes 0.5 0.0 0.0

Tax on bank accounttransactions (CPMF)

No 0.3 0.0 0.0

Fund for improvement ofauditing (FUNDAF)

No 0.3 0.0 0.0

StateTax on movement of

goods and services(ICMS)

Yes 229.4 20.9 7.2

Others No 36.9 3.4 1.2Municipal

Tax on services (ISS) Yes 31.1 2.9 1.0Real estate tax (IPTU) Yes 13.3 1.2 0.4




detailed information about many labor and nonlabor income sources, direct

taxes paid, transfers received, use of public education, and consumption.

When POF does not include questions on certain items (such as the amount

of consumption taxes paid), the values are allocated following the meth-

odologies described later. Data on the use of public health services come

from the Pesquisa Nacional por Amostra de Domicılios (PNAD; National

Household Sample Survey), 2008, which contains income data and a

detailed supplemental health survey containing the necessary information

regarding the use of public health services. PNAD 2008 has national cover-

age, sampling 118,138 households using a three-stage stratified sample

design. The main survey is conducted annually except in census years; the

health supplement, however, was only conducted in 2003 and 2008.7 Both

POF and PNAD are representative at the state level. Data on government

revenues and spending, which are used to scale-up household survey data

for the inequality (but not poverty) calculations, come from Brazil’s

national accounts and administrative data.8

In general, the amounts received from direct transfers are directly iden-

tified from the survey. The number of Bolsa Famılia beneficiaries captured

by the survey (7.3 million households), however, is significantly lower than

the number reported in national accounts (12.4 million households).9 Souza

(2010) shows that much of this discrepancy can be attributed to the survey’s

sample design. To correct for this problem, we use a propensity score

matching method suggested by Souza, Osorio, and Soares (2011) to impute

benefits to the missing 5.1 million households, selecting households that are

Table 2. (continued)

Included inanalysis

Billions ofReais

Percentage oftotal

Percentage ofGDP

ContributionsContributions to federal

pension fundsYes 200.7 18.3 6.3

Contributions to statepension funds

Yes 20.3 1.9 0.6

Contributions tomunicipal pension funds

Yes 5.6 0.5 0.2

Total Part 1,096.5 100.0 34.4

Sources: Amaral et al. (2011); Ministerio da Fazenda, Ministerio do Trabalho, and Ministerio daPrevidencia e Assistencia Social.Note: GDP ¼ gross domestic product.




very similar to beneficiary households but did not report receiving benefits.

After applying this method, both the number of beneficiaries and the total

program benefits in our data approximate the corresponding amounts in

national accounts. This method has very little impact on inequality results,

increases the poverty-reducing impact of direct transfers, and increases the

coverage of direct transfers among the poor, compared to the results when

the method is not implemented.10 Milk transfers from PAA Leite are

inferred: households in the Northeast and nonmetropolitan Minas Gerais

who reported the milk they consumed as being donated are assumed to have

received that milk from the program. Using this method, total benefits from

the program are slightly lower than in national accounts (the ratio of

national accounts to the survey total is 1.09).

On the tax side, individual income taxes (imposto de renda pessoa fısica

[IRPF] and the portion of ISS paid by workers) and property taxes (Imposto

Predial Territorial Urbano [IPTU] and Imposto Territorial Rural [ITR]) are

directly identified in the survey.11 By using the values reported in the sur-

vey, we are implicitly assuming that the incidence of the individual income

tax is borne entirely by labor (i.e., those workers who report paying it in the

household survey) and property taxes entirely by the owners of property

(i.e., those who report them in the survey). The payroll tax paid by employ-

ers Fundo de Garantiapor Tempo de Servico (FGTS) is not included in the

survey, but we simulate it, assuming that its burden is borne fully by work-

ers who had some deduction (individual income tax, contribution to the

pension system, or other deduction) from their labor income. Under this

assumption, reported labor incomes are net of FGTS, so we construct a new

pre-FGTS labor income counterfactual for formal sector workers and use

this variable in the market income aggregate.

Consumption taxes are imputed by applying effective tax rates to the very

detailed consumption data available from the survey. For ICMS and IPI, con-

sumption goods are grouped into nine categories: food, alcoholic beverages

and tobacco, clothing, electricity and domestic fuel, housing, health and edu-

cation, transport and communication, recreation and culture, and other goods

and services. We then multiply the amount spent in these categories by the

average effective tax rates calculated by Nogueira, Siqueira, and Souza

(2011) for each category. For PIS and COFINS, we use effective tax rates

by decile calculated by Rezende and Afonso (2010). Implicitly, we are

assuming that the incidence of consumption taxes falls fully on consumers.

To impute indirect subsidies, we use the total spent on electricity, in

combination with income, to determine who was eligible for the electricity

subsidy. Since we have data on the total spent on electricity but not total




consumption in kilowatt hours, we work backward to arrive at total con-

sumption by collecting data on the electricity rates of all Brazilian electricity

companies. Within each state, we average electricity rates across companies

and incorporate the tax laws for electricity in that state in order to determine

the amounts of spending that correspond to each energy consumption bracket.

We assume that all eligible households received the subsidy. The figure for

total spending on TSEE in national accounts is 1.3 times higher than total

benefits using this method, which is likely a result of leakages.

In-kind education benefits are equal to the average spending per student

by level (early childhood development, preschool, primary, lower second-

ary, upper secondary, and tertiary), which is obtained from national

accounts and imputed to students who attend public school. To estimate

in-kind health benefits, we take advantage of the supplement to the 2008

PNAD survey, which asks detailed questions relating to the use of health

services. We first group the types of health services reported in PNAD into

the three aggregate categories for which we have spending by state in

national accounts: primary care, inpatient care, and preventative care. Then,

for each of Brazil’s twenty-six states plus the federal district, we calculate

the average benefit received per health facility visit by dividing the total

spent in that state (combining spending at the federal, state, and municipal

levels) on that type of care by the total number of patient visits in the past

year when the patient received that particular type of care. Then the values

obtained for these benefits, which vary by state and type of care (primary,

inpatient, or preventative), are imputed to the households that report attend-

ing a public health facility and receiving that type of health service from the

SUS (essentially, from a free public health facility). Our method closely fol-

lows the best practices for health benefit incidence analysis outlined in

O’Donnell et al. (2008). We calculate the concentration coefficients of each

type of health care directly in the PNAD data set. To generate final income,

however, we impute health benefits back into the POF data, using the aver-

age benefit in PNAD by vintile (group of 5 percent of the population).

Results

To assess the impact of taxes and social spending, we use a variety of mea-

sures of inequality and poverty, the concentration of benefits received and

taxes paid with respect to market income, and effectiveness indicators.12

Market income inequality is very high in Brazil, with a Gini coefficient

of 0.58 (table 3). Through direct taxes and transfers, Brazil is able to reduce

inequality by 6 percent, which is impressive by Latin American but not




Western European standards.13 Spending on highly redistributive programs

is low, while much less redistributive programs are larger. Through all taxes

and transfers (direct and indirect taxes, direct and in-kind transfers, and

indirect subsidies), Brazil reduces inequality by 24 percent. The effective-

ness indicators for direct transfers and all transfers, respectively, are 0.90

and 1.67 in the benchmark scenario. These indicate relatively low

Table 3. Gini and Headcount Index for Different Income Concepts, Brazil 2009.

Marketincome

Net marketincome

Disposableincome

Post-fiscalincome

Finalincome

Benchmark scenarioGini 0.579 0.565 0.544 0.546 0.439

Headcount index (%)US$1.25 PPP/day (%) 5.8 5.9 2.7 4.4 —US$2.50 PPP/day (%) 15.1 15.7 11.2 16.3 —US$4.00 PPP/day (%) 26.2 27.2 23.2 31.0 —70 Reais per month (%) 6.4 6.6 3.1 5.2 —140 Reais per month (%) 16.5 17.1 12.7 18.2 —

Sensitivity analysis 1: contributory pensions as a government transfer

Gini 0.600 0.594 0.541 0.543 0.434


Sensitivity analysis 2: special pensions and contributory pensions as market income

Gini 0.573 0.559 0.544 0.546 0.439


Source: Authors’ calculations based on Pesquisa de Orcamentos Familiares, 2008�2009.Note: PPP ¼ purchasing power parity.




effectiveness: Brazil’s direct transfers effectiveness indicator is the lowest

among all countries in this issue.

To measure the impact of fiscal policy on poverty in a middle-income

country, we use the international poverty lines proposed by the World Bank

of US$1.25 PPP per day (ultra poverty), US$2.50 PPP per day (extreme

poverty), and US$4.00 PPP per day (moderate poverty), as well as the lines

used to determine eligibility for Bolsa Famılia’s fixed benefit (70 Reais per

month [US$1.34 PPP per day]) and variable benefit (140 Reais per month

[US$2.69 PPP per day]). Ultra poverty is reduced by 54 percent by direct

transfers (net of any direct taxes paid), extreme poverty by 26 percent, and

moderate poverty by just 11 percent. However, when indirect taxes are con-

sidered, the reduction in ultra poverty is significantly tempered, and

extreme and moderate poverty actually increase when one compares market

income with post-fiscal income. In other words, the number of near-poor

who are pushed into moderate poverty by paying more in taxes than they

receive in benefits (i.e., direct transfers and indirect subsidies) is higher than

the number of poor who escape poverty by receiving more in transfers and

subsidies than they pay in taxes.

The moderate success of direct transfers at reducing poverty can be

attributed to high coverage of the poor: 85 percent of the poor live in house-

holds receiving at least one direct transfer. The fact that poverty is not

reduced further despite Brazil’s high spending on direct transfers is due

to high leakages to the nonpoor (in addition to the deleterious effect of indi-

rect taxes): 73 percent of total direct transfer benefits go to the nonpoor. As

a result, the amount remaining to transfer to the poor is spread thinly: the

average transfer size of Bolsa Famılia, for example, is just US$0.35 PPP per

day in household per capita terms.

Table 4 shows concentration coefficients and budget sizes for direct

transfer programs, contributory pensions, energy subsidies, education and

health spending, and overall social spending. Bolsa Famılia, BPC, and milk

transfers are well targeted to the poor, with concentration coefficients of

�0.58, �0.48, and �0.36, respectively. However, unemployment benefits,

special circumstances pensions, scholarships, and other direct transfers

are progressive only in relative terms (i.e., their concentration curves

with respect to market income lie everywhere between the market income

Lorenz curve and the 45� line and, thus, are equalizing). As a result of these

opposing forces, the concentration curve of direct transfers as defined in

the benchmark case crosses the 45� line (figure 1), implying that they are

progressive, but not everywhere progressive in absolute terms. The curve

is initially concave and above the 45� line; the bottom three deciles receive




Tab

le4.

Conce

ntr

atio

nC

oef

ficie

nts

and

Budge

tSi

zes

for

Sele

cted

Pro

gram

s,Bra

zil2009.a

Pro

gram

Conce

ntr

atio

nco

effic

ient

with

resp

ect

toben

chm

ark

case

mar

ket

inco

me

Conce

ntr

atio

nco

effic

ient

with

resp

ect

tose

nsi

tivi

tyan

alys

is1

mar

ket

inco

me

Conce

ntr

atio

nco

effic

ient

with

resp

ect

tose

nsi

tivi

tyan

alys

is2

mar

ket

inco

me

Budge

tsi

ze(p

erce

nta

geofG

DP)

Spec

ialci

rcum

stan

ces

pen

sions

0.2

00.0

4—

2.2

8U

nem

plo

ymen

tben

efits

0.1

80.2

50.1

70.5

8BPC

(nonco

ntr

ibuto

rypen

sions)

�0.4

8�

0.4

9�

0.4

80.5

3Bols

aFa

mıli

a(C

CT

)�

0.5

8�

0.5

1�

0.5

90.3

9O

ther

dir

ect

tran

sfer

sb0.1

50.2

10.1

50.2

6Sc

hola

rship

s0.2

80.3

10.2

80.1

1M

ilktr

ansf

erpro

gram

�0.3

5�

0.3

3�

0.3

60.0

1D

irec

ttr

ansf

ers

excl

udin

gsp

ecia

lci

rcum

stan

ces

pen

sions

�0.2

2�

0.1

8�

0.2

31.8

7

Dir

ect

tran

sfer

sin

cludin

gsp

ecia

lci

rcum

stan

ces

pen

sions

0.0

3�

0.0

5—

4.1

6

Contr

ibuto

rypen

sions

—0.0

6—

9.0

6D

irec

ttr

ansf

ers

plu

sco

ntr

ibuto

rypen

sions

—0.0

2—

13.2

1

Pre

school

�0.3

3�

0.2

5�

0.3

40.3

0Pri

mar

yed

uca

tion

�0.3

1�

0.2

5�

0.3

22.3

6Se

condar

yed

uca

tion

�0.2

1�

0.1

6�

0.2

20.3

8T

ertiar

yed

uca

tion

0.4

40.4

20.4

40.8

2T

ota

led

uca

tion

spen

din

g�

0.1

5�

0.1

1�

0.1

65.3

1

360



Tab

le4.

(continued

)

Pro

gram

Conce

ntr

atio

nco

effic

ient

with

resp

ect

toben

chm

ark

case

mar

ket

inco

me

Conce

ntr

atio

nco

effic

ient

with

resp

ect

tose

nsi

tivi

tyan

alys

is1

mar

ket

inco

me

Conce

ntr

atio

nco

effic

ient

with

resp

ect

tose

nsi

tivi

tyan

alys

is2

mar

ket

inco

me

Budge

tsi

ze(p

erce

nta

geofG

DP)

Pri

mar

yca

re�

0.1

6�

0.1

1�

0.1

61.0

5In

pat

ient

care

�0.1

1�

0.1

6�

0.0

92.5

6Pre

venta

tive

care

�0.1

5�

0.1

9�

0.1

30.2

9T

ota

lhea

lth

spen

din

g�

0.1

1�

0.1

6�

0.1

05.2

1Ener

gysu

bsi

die

s�

0.2

7�

0.3

0�

0.2

70.0

5So

cial

spen

din

gex

cludin

gsp

ecia

lci

rcum

stan

ces

pen

sions

�0.1

5�

0.1

1�

0.1

513.8

9

Soci

alsp

endin

gin

cludin

gsp

ecia

lci

rcum

stan

ces

pen

sions

�0.0

9�

0.0

9—

16.1

7

Soci

alsp

endin

gplu

sco

ntr

ibuto

rypen

sions

—�

0.0

4—

25.2

3

Not

e:BPC¼

Ben

efıc

iode

Pre

stac

aoC

ontinuad

a;C

CT¼

conditio

nal

cash

tran

sfer

;G

DP¼

gross

dom

estic

pro

duct

.So

urce

:A

uth

ors

’ca

lcula

tions

bas

edon

Pes

quis

ade

Orc

amen

tos

Fam

iliar

es,2008�

2009.

a All

conce

ntr

atio

nco

effic

ients

are

stat

istica

llysi

gnifi

cant

from

zero

atth

e1

per

cent

sign

ifica

nce

leve

l.T

he

table

incl

udin

gst

andar

der

rors

isav

aila

ble

from

the

auth

ors

upon

reques

t.bO

ther

dir

ect

tran

sfer

sin

clude

assi

stan

cefr

om

PIS

/PA

SEP,B

ols

aEsc

ola

,Auxıli

oG

as,o

ther

auxıli

os,

child

labor

era-

dic

atio

n,m

inim

um

inco

me

pro

gram

s,an

dth

ebas

icfo

od

bas

ket

pro

gram

.

361



a larger share of direct transfers than their population share due to highly

progressive programs like Bolsa Famılia. However, a large chunk of trans-

fers from other programs (relative to population shares) is concentrated at

the top of the distribution.

Indirect subsidies are progressive in absolute terms, with a concentration

coefficient of �0.27. Education spending is progressive in absolute terms

overall; its only component that is not progressive in absolute terms is ter-

tiary education. It is worth noting that the concentration coefficient of ter-

tiary education, at 0.43, makes Brazil one of the worst performers in Latin

America in terms of providing tertiary education access to the poor. Health

spending and all of its components are progressive in absolute terms. Over-

all social spending is progressive in absolute terms—this is a robust result

that holds for the different definitions of social spending that arise in the

benchmark scenario and sensitivity analyses, as shown in table 4.

Conclusion

We calculate the effects of fiscal policy on income distribution and poverty

in Brazil. In terms of direct transfers, Brazil has relatively high spending

and low effectiveness. Bolsa Famılia, BPC, and milk transfers are well tar-

geted to the poor and highly progressive in absolute terms, but other much

Figure 1. Concentration curves with respect to market income (benchmarkanalysis), Brazil 2009.




larger direct transfers are progressive only in relative terms. Brazil is also a

relatively high spender on health and education compared to the other coun-

tries studied in this special issue. With the exception of tertiary education,

all components of public health and education spending are progressive in

absolute terms. On the tax side, there is a substantial deleterious effect of

indirect taxes on poverty. In many cases, the benefits of transfer programs

and indirect subsidies are offset by indirect taxes. A reform of the indirect

tax system—especially with respect to taxes on basic food items—or larger,

well-targeted compensating transfers to offset the costs of indirect taxes for

the poor must be a high priority.

Acknowledgment

The authors are grateful to Nora Lustig, Francisco Ferreira, Otaviano Canuto, par-

ticipants of the Commitment to Equity Workshop (hosted by the Inter-American

Dialogue and CIPR, November 2011) and the Workshop on Fiscal Incidence in LAC

(hosted by the World Bank, May 2012), Nadia von Jacobi, two anonymous referees,

and the editor for excellent feedback on earlier versions of this research; to Sergei

Soares for providing the code to implement his methodology; to Adam Ratzlaff for

research assistance; and to Emily Travis and Scott Odell for copyediting. We are

also grateful to the Center for Inter-American Policy and Research, Inter-

American Development Bank, Inter-American Dialogue, International Fund for

Agricultural Development, Latin American Development Bank, and World Bank

for funding for related research that contributed to the development of this article.

Authors’ Note

Any errors are the responsibility of the authors.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research,

authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research,

authorship, and/or publication of this article:The authors received financial support

for related research that contributed to the development of this article from the Cen-

ter for Inter-American Policy and Research, Inter-American Development Bank,

Inter-American Dialogue, International Fund for Agricultural Development, Latin

American Development Bank, and World Bank.




Notes

1. This observation is a result of authors’ calculations using microdata from the

Pesquisa Nacional por Amostra de Domicılios (PNAD); the result holds regard-

less of whether inequality is measured by the Gini, mean log deviation, or

Theil’s T index. Furthermore, following Atkinson (1969), we can state that

inequality was unambiguously lower in 2011 than in 2001 using any inequality

measure that reflects a social welfare function that is an additively separable and

symmetric function of individual incomes, since the Lorenz curve for 2011

dominates that of 2001 and mean real income was higher in 2011.

2. Headcount rates over the period 1995 to 2011 are available from Instituto de

Pesquisa Economica Aplicada (IPEA;2012) and poverty gap and squared pov-

erty gap indices over the period 1981 to 2009 from Socioeconomic Database for

Latin America and the Caribbean (sedlac.econo.unlp.edu.ar).

3. Although individuals’ own reports of the income taxes they paid will be imper-

fect, we prefer this method to microsimulation because it captures informality

and evasion more accurately than microsimulations, even when the latter are

well designed. Recent incidence analyses for Brazil using microsimulation

include Immervoll et al. (2009) and Nogueira, Siqueira, and Souza (2011). See

Soares et al. (2009) for a comparison of individual income taxes reported in Pes-

quisa de Orcamentos Familiares to those resulting from a microsimulation

model. It is also worth noting that the rates of individual income tax underre-

porting and labor income underreporting (when the totals for these categories

in our analysis are compared to the totals in national accounts) are almost iden-

tical, suggesting that taxes are not misreported to a greater extent than incomes.

4. The number of individuals in beneficiary households is according to the Ministry of

Social Development (MDS). For more information about Bolsa Famılia, Soares

(2012) provides an excellent overview of its history, design features, and impact.

5. According to the MDS, see http://aplicacoes.mds.gov.br/sagi/miv/miv.php.

6. As explained in Lustig, Pessino, and Scott (2014), there is no agreement in the

literature on whether contributory pensions should be treated as part of market

income like income from savings or as a government transfer.

7. For a comparison of the distribution of income found in the two surveys, see

Barros, Cury, and Ulyssea (2007).

8. For the precise sources for each component of social spending and revenues, see

tables 1 and 2, respectively.

9. The number of transfer beneficiaries reported by Brazil’s MDS is regarded to be

accurate. Beneficiaries are part of the Cadastro Unico (Single Registry) database;

information in the database is collected by municipal agents, then sent to the

MDS and Caixa Economica Federal, the government-controlled bank responsible




for transfer payments. Thus, the number of beneficiaries reported by MDS

matches those receiving payments through Caixa Economica Federal. In addition,

the list of Bolsa Famılia recipients is part of an open access database.

10. Results without this method are available from the authors upon request.

11. We follow Silveira et al. (2011) and Rezende and Afonso (2010) in assuming—

based on the survey’s interviewer manual—that the portion of ISS paid by

workers is captured by the labor income question on other deductions.

12. The effectiveness indicators measure how well governments reduce inequality

and poverty per amount spent. The inequality reduction effectiveness indicator

for direct transfers is defined as the proportional change between the net market

and disposable income Ginis (which can be exclusively attributed to direct

transfers) divided by the amount spent on direct transfers as a percentage of

gross domestic product (GDP).

13. Direct taxes and transfers reduce inequality by about one-third on average in

Europe (Immervoll et al. 2006). Our results are consistent with Immervoll

et al. (2009), who demonstrate the limited redistributive effects of fiscal policy

in Brazil (using a different data set and microsimulations), despite its high level

of taxation (35 percent of GDP) and high spending on social programs.

References

Amaral, Gilberto Luiz, Joao Eloi Olenike, and Letıcia Mary Fernandes do Amaral

Viggiano. 2007. ‘‘Estudo sobre o verdadeiro custo da tributacao brasileira.’’ Insti-

tuto Brasileiro de Planejamento Tributario Working Paper, Instituto Brasileiro de

Planejamento Tributario, Sao Paulo, Brazil.

Amaral, Gilberto Luiz, Joao Eloi Olenike, Letıcia Mary Fernandes do Amaral Vig-

giano, and Fernando Steinbruch. 2011. ‘‘Carga tributaria brasileira de 2010.’’ Insti-

tuto Brasileiro de Planejamento Tributario Working Paper, Instituto Brasileiro de

Planejamento Tributario, Sao Paulo, Brazil.

Atkinson, Anthony B. 1969. ‘‘On the Measurement of Inequality.’’ Journal of Eco-

nomic Theory 2:244–63.

Barros, Ricardo, Mirela de Carvalho, Samuel Franco, and Rosane Mendonca. 2010.

‘‘Markets, the State, and the Dynamics of Inequality in Brazil.’’ In Declining

Inequality in Latin America: A Decade of Progress?, edited by Luis F. Lopez-

Calva and Nora Lustig, 134–74. Washington, DC: Brookings Institution.

Barros, Ricardo, Samir Cury, and Gabriel Ulyssea. 2007. ‘‘A Desigualdade no Brasil

Econtra-Se Subestimada?’’ In Desigualdade de Renda no Brasil: Uma Analise

da Queda Recente, edited by Ricardo Paes de Barros, Miguel Nathan Foguel, and

Gabriel Ulyssea, 237–73. Brasılia, Brazil: IPEA.

Corbacho, Ana, Vicente Fretes Cibils, and Eduardo Lora. eds. 2013. More than

Revenue: Taxation as a Development Tool. New York: Palgrave Macmillan.




Ferreira, Francisco H. G., Phillippe G. Leite, and Julie A. Litchfield. 2008. ‘‘The

Rise and Fall of Brazilian Inequality: 1981–2004.’’ Macroeconomic Dynamics

12:199–230.

Gasparini, Leonardo, and Nora Lustig. 2011. ‘‘The Rise and Fall of Income Inequal-

ity in Latin America.’’ In The Oxford Handbook of Latin American Economics,

edited by Jose Antonio Ocampo and Jaime Ros, 691–714. New York: Oxford

University Press.

Higgins, Sean. 2012. ‘‘The Impact of Bolsa Famılia on Poverty: Does Brazil’s

Conditional Cash Transfer Program have a Rural Bias?’’ Journal of Politics

and Society 23:88–125.

Immervoll, Herwig, Horacio Levy, Christine Lietz, Daniela Mantovani, Cathal

O’Donoghue, Holly Sutherland, and Gerlinde Verbist. 2006. ‘‘Household

Incomes and Redistribution in the European Union: Quantifying the Equalizing

Properties of Taxes and Benefits.’’ In The Distributional Effects of Government

Spending and Taxation, edited by Dimitri, B. Papadimitriou, 135–65. New York:

Palgrave Macmillan.

Immervoll, Herwig, Horacio Levy, Jose Ricardo Nogueira, Cathal O’Donoghue, and

Rozane Bezerra de Siqueira. 2009. ‘‘The Impact of Brazil’s Tax-Benefit System

on Inequality and Poverty.’’ In Poverty, Inequality, and Policy in Latin America,

edited by Stephan Klasen and Felicitas Nowak-Lehmann, 271–301. Cambridge,

MA: MIT Press.

INSS (Instituto Nacional de Seguro Social). 2010. ‘‘Regime geral de previdencia

social: Balanco do ano de 2009.’’ Informe de Previdencia Social 22:1–16.

IPEA (Instituto de Pesquisa Economica Aplicada). 2012. ‘‘A decada inclusiva

(2001-2011): Desigualdade, pobreza e polıticas de renda.’’ Comunicacao do

IPEA No. 155, IPEA, Brasılia, Brazil.

Lustig, Nora, Luis Felipe Lopez-Calva, and Eduardo Ortiz-Juarez. 2013. ‘‘Declining

Inequality in Latin America in the 2000s: The Cases of Argentina, Brazil, and

Mexico.’’ World Development 44:129–41.

Lustig, Nora, Carola Pessino, and John Scott. 2014. ‘‘The Impact of Taxes and Social

Spending on Inequality and Poverty in Argentina, Bolivia, Brazil, Mexico, Peru and

Uruguay: Introduction to the Special Issue.’’ Public Finance Review 42(3):287–303.

Ministerio do, Trabalho. 2011. Relatorio de avaliacao do plano plurianual,

2008–2011. Brasılia, Brazil: Ministerio do Trabalho.

Nogueira, Jose Ricardo Bezerra, Rozane Bezerra de Siqueira, and Evaldo Santana

de Souza. 2011. ‘‘A Brazilian Tax-Benefit Microsimulation Model.’’ In Micro-

simulation Models for Latin America, edited by Carlos M. Urzua, 19–50. Mexico

City: ITESM.

O’Donnell, Owen, Eddy van Doorslaer, Adam Wagstaff, and Magnus Lindelow.

2008. ‘‘Analyzing Health Equity Using Household Survey Data: A Guide to




Techniques and Their Implementation.’’ WBI Learning Resources Series, The

World Bank, Washington, DC.

Portal, Tributario. 2012. Os tributos no Brasil. http://www.portaltributario.com.br/

tributos.htm. Accessed September 20, 2012.

Rezende, Fernando, and Jose Roberto Afonso. 2010. ‘‘Equidade Fiscal no Brasil.’’ In

Equidad Fiscal en Brasil, Chile, Paraguay y Uruguay, edited by A. Barreix, L. Vil-

lela, and J. Roca, 34–105. Washington, DC: Inter-American Development Bank.

Silveira, Fernando Gaiger, Johnatan Ferreira, Joana Mostafa, and Jose A. Carlos

Ribeiro. 2011. ‘‘Qual o impacto da tributacao e dos gastos publicos sociais na

distribuicao de renda no Brasil?’’ In Progressividade da Tributacao e Desoner-

acao da Folha de Pagamentos, edited by Jose A. Carlos Ribeiro, Alvaro

Luchiezi Jr., and Sergio E. Arbulu Mendonca. 25–64. Brasılia, Brazil: IPEA.

Siqueira, Rozane Bezerra, Jose Ricardo Bezerra Nogueira, and Evaldo Santana de

Souza. 2010. ‘‘Aliquotas efetivas e a distribuicao da carga tributaria indireta

entre as famılias no Brasil.’’ XV Premio do Tesouro Nacional 2010, Secretaria

do Tesouro Nacional, Brasılia, Brazil.

Soares, Sergei. 2012. ‘‘Bolsa Famılia, Its Design, Its Impacts and Possibilities for

the Future.’’ International Policy Centre for Inclusive Growth Working Paper,

International Policy Centre for Inclusive Growth, Brasılia, Brazil.

Soares, Sergei, F. G. Silveira, C. H. Dos Santos, F. M. Vaz, and A. L. Souza. 2009.

‘‘O potencial distributivo do imposto de renda pessoa fısica (IRPF).’’ IPEA

Working Paper, IPEA, Brasılia, Brazil.

Souza, Pedro H. G. F. 2010. ‘‘Uma metodologia para decompor diferencas entre

dados administrativos e pesquisas amostrais, com aplicacao para o Programa

Bolsa Famılia e o Benefıcio de Prestacao Continuada na PNAD.’’ IPEA Working

Paper, IPEA, Brasılia, Brazil.

Souza, Pedro H. G. F., Rafael G. Osorio, and Sergei Soares. 2011. ‘‘Uma metodologia

para simular o Programa Bolsa Familia.’’ IPEA Working Paper, IPEA, Brasılia,

Brazil.

Author Biographies

Sean Higgins is a PhD student in the Department of Economics at Tulane Univer-

sity. His research has focused on the impact of taxes and cash transfer programs on

poverty in Brazil.

Claudiney Pereira is a senior professor of practice in the Department of Economics

at Tulane University. He received his doctorate in economics from North Carolina

State University. His research has focused on the empirics of economic growth,

monetary policy and the role of the financial sector in Brazil, and fiscal policy

effects on poverty and income distribution in Brazil.



http://www.portaltributario.com.br/tributos.htm

http://www.portaltributario.com.br/tributos.htm


the effects of brazil's taxation and social spending on the distribution of household income

Documents