Poverty, Inequality and Social Policies in Brazil: 1995-2009

POVERTY, INEQUALITY AND SOCIAL
POLICIES IN BRAZIL, 1995-2009

Working Paper number 87 February, 2012

Pedro H. G. Ferreira de Souza
Institute for Applied Economic Research (IPEA)

International

Centre for Inclusive Growth

Copyright© 2012
International Policy Centre for Inclusive Growth
United Nations Development Programme

International Policy Centre for Inclusive Growth (IPC - IG)
Poverty Practice, Bureau for Development Policy, UNDP
Esplanada dos Ministérios, Bloco O, 7º andar
70052-900 Brasilia, DF - Brazil
Telephone: +55 61 2105 5000

E-mail: ipc@ipc-undp.org URL: www.ipc-undp.org

The International Policy Centre for Inclusive Growth is jointly supported by the Poverty Practice,
Bureau for Development Policy, UNDP and the Government of Brazil.

Rights and Permissions

All rights reserved.

The text and data in this publication may be reproduced as long as the source is cited.
Reproductions for commercial purposes are forbidden.

The International Policy Centre for Inclusive Growth disseminates the findings of its work in
progress to encourage the exchange of ideas about development issues. The papers are
signed by the authors and should be cited accordingly. The findings, interpretations, and
conclusions that they express are those of the authors and not necessarily those of the
United Nations Development Programme or the Government of Brazil.

Working Papers are available online at www.ipc-undp.org and subscriptions can be requested
by email to ipc@ipc-undp.org

Print ISSN: 1812-108X

POVERTY, INEQUALITY AND SOCIAL POLICIES IN BRAZIL,
1995-2009

Pedro H. G. Ferreira de Souza *

ABSTRACT
Since the mid-1990s, Brazil has undergone extensive reforms that have finally reversed the
dismaying economic performance of the 1980s. In particular, poverty and inequality indicators
have improved dramatically, especially since the late-2000s. This paper provides an overview
of such recent trends and discusses the role played by four major government interventions:
public education, the minimum wage law, Social Security pensions and Social Assistance
transfers. Additionally, available data sets and methods for policy evaluation are also discussed.

1 INTRODUCTION
By the end of the first decade of the 21st century, the most usual international depiction of
Brazil is that of a burgeoning, upcoming country. Although in many ways frankly exaggerated,
this marks a stark contrast with a not-so-distant past. Back in the early 1990s, such optimistic
prospects were almost unimaginable by most Brazilians—and rightly so. This turnaround has
had a lot to do with favourable international circumstances during much of the 2000s, but they
also owe a lot to extensive reforms that made possible something that was almost
unprecedented in Brazil throughout the past century: sustained pro-poor growth.
This paper aims to place in context and summarize some of these transformations,
focusing specifically on the evolution and assessment of poverty and inequality trends as well
as of the major government interventions in these areas. This is not unwarranted, as some of
the most important transformations of the past decade and a half are closely connected to an
improved and increasingly effective social protection network that, as we shall see,
contributed decisively to the decline in poverty and inequality.
As the currently existing institutions in these areas are obviously still far from perfect,
there is a lot to like, but also a lot to dislike about them. Hence, this paper provides an overview
of the most important criticisms levelled at them.
Finally, poverty and inequality research is a technical field of expertise that is
heavily dependent upon high-quality data and proper research tools. One of this paper’s
secondary goals, then, is to briefly describe the most used data sets and the most prevalent
methodologies used by Brazilian researchers to monitor poverty and inequality trends and to
assess the effects of those major social programmes.
The rest of the paper is organized as follows: the second section provides an overview of
the Brazilian macro-economic context since the late 1980s; the third section describes more
thoroughly the most important data sets available; the fourth section reports on the overall

* Institute for Applied Economic Research (IPEA).

2 International Policy Centre for Inclusive Growth

income, poverty and inequality trends in Brazil since the mid-1990s; the fifth section offers an
in-depth look at four major governmental interventions, namely, educational and minimum
wage policies, Social Security and Social Assistance; the sixth section shows the results
for the most common decompositions of income inequality in order to identify the major
determinants of the swift decline observed between 2001 and 2009; finally, the seventh
section summarizes the most important findings.

2 MACRO-ECONOMIC CONTEXT1
After a stint of high GDP growth rates in the late 1960s and early 1970s—an event dubbed the
‘Economic Miracle’—the Brazilian economy entered a bleak period as it tried to grapple with
the balance of payments crisis caused by that decade’s oil shocks and mounting external debt.
Between 1986 and 1991, no fewer than five economic plans were enacted in order to rein in
spiralling inflation rates and to boost the erratic economic growth. None fully succeeded and,
worse, the economic crisis became entangled with a political one as Fernando Collor de Mello,
the first elected president since the 1964 military coup, resigned in 1992 in an attempt to stop
his impeachment trial on charges of corruption.
By then, Brazil was in turmoil. During the 1980s, both poverty and inequality went up, but
the downfall of the military regime, the new Federal Constitution of 1988, and free direct
presidential elections had been celebrated as reasons for hope. In the early 1990s, all that was
already starting to fade, as the economy was in a shambles—in 1992, the annual inflation rate
reached almost 1,150 per cent and the GDP contracted by 0.5 per cent —and the political crisis
continued after Collor’s departure.
In that context, 1994’s Plano Real (Real Plan) was a real turning point. The plan introduced
a new currency (the real) and sought to bring inflation under control through a combination
of extensive trade liberalization, privatizations, contractionary monetary policies and high
interest rates in order to attract foreign capital and finance the current accounts deficit.
Even though the plan was successful—annual inflation rates dropped from a staggering
2,500 per cent in 1993 to 930 per cent in 1994 and to just 22 per cent in 1995—the lack of a
serious fiscal adjustment resulted in an increased public sector debt and overvalued currency
rates, which were necessary to keep demand-side inflation low. Hence, the Asian and Russian
crises of 1997 and 1998 set off a currency crisis in Brazil, as the real suffered a severe devaluation.
This led to another round of reforms, as the Central Bank eschewed the fixed-exchange
rates system and adopted inflation-targeting policies. Also, deficit-cutting policies were enacted,
which entailed a period of very low growth rates: between 1994 and 1997, immediately after the
Plano Real, GDP growth averaged 3.8 per cent per year; between 1998 and 2003, the macro-
economic woes dragged down the average growth rates to just 1.6 per cent per year.
Still, the basic conditions—both macro-economic and institutional—were set for
renewed economic growth. The circumstances became favourable in the mid-2000s, as the
rapid industrialization of China and other developing countries resulted in rising commodity
prices, which favourably affected Brazilian exports. Additionally, abundant international credit
also increased foreign direct investment.
Overall, this created a new opportunity window that was seized by the Brazilian
government with expansionary fiscal policies to jumpstart the economy. In particular,

Working Paper 3

there was an aggressive expansion of both consumer and corporate credit and increased
spending on social policies, such as Social Security pensions and Social Assistance transfers.
The end result was a robust consumer-led economic boom with an average GDP growth
rate of approximately 4.4 per cent per year between 2004 and 2010, the highest since the late-
1970s.2 Moreover, even though there was a 0.6 per cent contraction in the wake of the global
financial crisis in 2009, GDP growth in 2010 rebounded to 7.5 per cent—the highest annual
growth rate in over two decades.
Such favourable conditions were reflected in a thriving labour market that promoted
vigorous formalization and job creation. Between 1999 and 2003, for instance, about
2.4 million formal jobs were created; between 2003 and 2007, this number more than
doubled, reaching 5.1 million; finally, between 2007 and 2011, it soared to 8.7 million.
Despite such promising performance, as the first decade of the 2000s came to an end,
there were renewed doubts regarding the sustainability of Brazilian growth. Low investment
rates, the precarious state of the infrastructure in transports, an overvalued currency, the threat
of rising prices, and glum prospects for the global economy have revived some analysts’
persistent suspicions that Brazil might just have benefitted from low-hanging fruit.
So far, this seems to be an inordinately pessimistic stance but, although the next few
years should see somewhat slower growth rates, the debate is still open. In terms of economic
fundamentals and institutional arrangements, Brazil is right now markedly different from the
1980s. The turmoil of the late 1980s seems long past, as the economic growth of the past
decade occurred under stable, democratic governments and brought about not only a sharp
reduction in poverty rates but also—for the first time in decades—a steady decline in income
inequality, as we shall see in the next sections.

3 DATA SETS
High-quality data sets are obviously the number-one requirement for poverty and inequality
tracking and policy evaluation. Usually, these are either household surveys with family-level
and individual-level information or administrative data. Even though the latter can be
potentially very useful, their use in such studies is still incipient, mostly due to confidentiality
clauses and their sheer size.
Fortunately, high-quality household surveys have been available for decades now in
Brazil. The first poverty and inequality studies were based on the 1960 and 1970 Population
Census, but the most widely used data set is the Pesquisa Nacional por Amostra de Domicílios
(PNAD), a national household survey conducted by IBGE, the central statistics office of the
country. The PNAD is a multi-purpose survey that was first fielded in 1967 and, after some
initial adjustments, settled on its current basic design in 1976. Since then, it has been carried
out yearly—except during census years (1980, 1991 and 2000) and in 1994—always in the last
quarter of the year. Although it covers several broad areas—such as education, migration,
fertility, income and earnings, housing, and household access to services and facilities—
a special emphasis is accorded to labour market participation.3
Other household surveys that are also used in poverty and inequality research are the
Pesquisa de Orçamentos Familiares (POF), which is an income and expenditures survey fielded
every five years (with limited coverage in 1987/1988 and 1995/1996 and full national coverage


in 2002/2003 and 2008/2009), the Pesquisa Mensal do Emprego (PME), a monthly labour
market-oriented survey that is restricted to metropolitan areas, and the National Population
Census, which is carried out every 10 years.
The PNADs’ popularity stems from some clear-cut advantages over those other surveys.
First, obviously other than the census, they have the largest sample sizes, roughly
equivalent to 100,000 households per year. Second, unlike the census, they have a long-
term series with minimum comparability issues, especially since the early 1980s, as most
changes have dealt with the survey’s coverage extension to far-flung areas of the country.
Third, they present generally very consistent and high-quality data. Fourth, they always
cover several broad areas, providing researchers with lots of variables to examine.
Finally, they permit fine-grained year-by-year analyses.
There are some drawbacks, though. First and foremost for poverty and inequality
researchers, it is known that incomes—especially property-related income and non-monetary
income—are underreported. This is a common issue of such surveys and affects mostly the very
rich and the very poor and thus does not seem to cause any serious bias in income inequality
measures (Barros, Cury & Ulyssea, 2006). Poverty indicators also remain somewhat unbiased if
one takes into account that the PNADs are mostly useful to measure monetary poverty.
Second, the PNADs are purely cross-sectional data sets. This is not a problem in itself,
but it must be stressed that there is a conspicuous lack of panel data and that is probably the
single biggest hurdle for poverty and inequality research in Brazil. Very little is known about life
trajectories, income volatility and such. Luckily, this seems to be about to change in the next
decade, as IBGE is updating its household surveys.
Finally, the PNADs’ three-stage sampling design is slightly biased toward larger
municipalities and therefore underestimates the scope of some social programmes, like the
Programa Bolsa Família, which has municipal quotas that favour the poorer and smaller towns
(Souza, 2010). It probably also underestimates the overall level of poverty and inequality.
None of these limitations offset the advantages. All in all, it is safe to assume that, even
though the PNADs are less than perfect when it comes to estimating the overall levels of
income, poverty and inequality, they are very useful to track trends and patterns.

4 TRENDS IN INCOME, POVERTY AND INEQUALITY IN BRAZIL
Unlike in many other countries, the standard income measurement unit in Brazil is the gross
household per capita income, after transfers but before taxes and without any equivalence
scales. Figure 1 presents the income evolution since the mid-1990s in Purchasing Power
Parity (PPP) US dollars4 (see World Bank, 2008). It shows clearly that 2003 marked an
inflection point: between 1995 and 2003, household per capita income grew just 1.3
per cent per year; afterward, between 2003 and 2009, it shot up more than 7 per cent per
year. The cumulative income growth throughout the entire 1995/2009 period reached
almost 70 per cent (3.8 per cent per year).
Due to methodology, the PPP factors slightly overstate recent Brazilian income growth.
Still, there has been substantial growth in the latter half of the past decade: per capita GDP
rose a meagre 3 per cent between 1995 and 2003, but 17 per cent between 2003 and 2009.

Working Paper 5

FIGURE 1
Mean Real Household Per Capita Income (US$ PPP) – Brazil, 1995/2009

Source: Pesquisa Nacional por Amostra de Domicílios, 1995-2009.
PPP factors from: World Bank. Global Purchasing Power Parities and Real Expenditures - 2005 International Comparison
Program. Updated by average annual inflation for Brazil and the United States: International Monetary Fund. World
Economic Outlook 2009.

As usual, this income growth was unevenly distributed. However, unlike in the 1960s
and 1970s, the gains this time accrued disproportionately to the bottom half of the income
distribution, as seen in Figure 2. Between 1995 and 2009, the income growth for the bottom
20 per cent reached 127 per cent and, for the top 20 per cent, just 54 per cent. Consequently,
the ratio between the average incomes of the top 20 per cent and the bottom 20 per cent
plummeted from 27 to 18 times—a whopping 32 per cent decline, although it is still high.

FIGURE 2
Real Income Growth by Centiles (%) – Brazil, 1995 and 2009

Source: Pesquisa Nacional por Amostra de Domicílios, 1995 and 2009.


The magnitude of such pro-poor growth was considerable, and poverty and inequality
indicators dropped noticeably. Figure 3 presents the evolution of extreme poverty as
measured by the World Bank’s and the Millennium Development Goal’s poverty line of
US$ PPP 1.25 per day. Poverty reduction dates back from the mid-1990s and picked up
speed in the mid-2000s as the economic recovery was combined with the fall in income
inequality. In 2007, Brazil achieved its own Millennium Development Goal, which was
to reduce poverty by 2015 to one quarter of its level in 1990 (25.6 per cent).

FIGURE 3
Extreme Poverty (US$ PPP 1.25/Day Poverty Line) – Brazil, 1995/2009


Poverty eradication was announced in 2011 as the top priority of the newly inaugurated
president, Dilma Rousseff, who set an official poverty line very similar to the World Bank’s.
Furthermore, the 2009 income distribution displays first-order dominance over both the 1995
and 2003 distributions. This means that, for any poverty line whatsoever, the relevant poverty
indicators should be lower in 2009 than in 1995 or 2003.
In Brazil, the most used synthetic inequality measure is the Gini index, as displayed in
Figure 4. As mentioned, after decades of stagnant or rising inequality, the good news of the
2000s was the swift decline in the Gini index, which dropped by 9 per cent between 2001 and
2009, reaching its lowest level since the 1970s (Barros et al., 2006). Remarkably, income
inequality began to fall before the economic recovery of the mid-2000s, highlighting the
importance of some longer-term social policies, as will be discussed further below.

Working Paper 7

FIGURE 4
Gini Index of the Household Per Capita Income – Brazil, 1995/2009


Nevertheless, Brazilian income inequality is still considerably large: even if the current
pace is maintained, it would take another couple of decades to reach the inequality levels
presently found in developed countries (Milanovic & Yitzhaki, 2002; Soares, 2008; López-Calva
& Lustig, 2010).
Likewise, the recent performance should not be mistaken for any kind of Brazilian
exceptionalism. The 2000s were a good decade for several developing economies, especially in
Latin America. Table 1, for instance, shows the recent growth and inequality trends in the six
largest Latin American countries. Brazil was not an outlier. All six countries went through a
period of pro-poor growth, with Peru standing out as the top performer in terms of economic
growth and inequality reduction.

TABLE 1
Annual GDP Growth and Variation in the GINI Index of Household Per Capita Income –
Six largest Latin American Countries, 2000s
GDP growth 2002‐2009 Change in the Gini index
Countries
(% per year) in the 2000s (%)
Argentina 3.7 ‐15
Brazil 3.7 ‐9
Chile 4.2 ‐6
Colombia 4.4 ‐1
Mexico 2.8 ‐6
Peru 5.6 ‐13
Venezuela 4.4 ‐1
Sources: GDP Growth: United Nations. World Economic Situation and Prospects 2011. Inequality: Socio-Economic
Database for Latin America and the Caribbean (CEDLAS and The World Bank). Note that, in order to ensure comparability,
CEDLAS makes a wide range of adjustments to the original data sets. The years used to estimate the Gini coefficient are as
follows: Argentina, 2003-2009; Brazil, 2001-2009; Chile, 2000-2009; Colombia, 2001-2004; Mexico, 2000-2008; Peru, 2003-
2009; Venezuela, 2000-2006.


5 POVERTY, INEQUALITY, AND THE STATE
Brazil has a long tradition of research on poverty and inequality and, as such, it is unsurprising
that the changes spurred by the last decade have drawn widespread attention. A vast array of
possible causes has been studied, dealing mostly with public policies and programmes.
Naturally, state interventions can impinge directly and indirectly on poverty and
inequality in a myriad of ways. Due to the popularity of the PNADs, some have been much
more studied than others and the rest of this paper will focus on them. That leaves out some
interventions—such as the expenditures on the Universal Health System and other public
services—the distributive effects of which are either ambiguous or hard to measure. More
important, it also leaves out some interventions that are known to be notoriously regressive,
like taxes. Previous research has shown that the Brazilian tax code, similar to that of other
developing countries, relies heavily on indirect consumption taxes, which take a greater
toll on the poor (Silveira, 2008).
In this section, four typical areas of intervention of the 20th century social policies will be
discussed in detail: public education policies, the minimum wage law, Social Security pensions,
and Social Assistance transfers.
The expenditures linked to those major areas are listed in Table 2 and comprised a hefty
16 per cent of GDP, or almost 50 per cent of the total tax revenues in 2006. As noted by many
authors (for instance, Rocha & Caetano, 2008), Social Security expenditures stand out in a
relatively young country like Brazil. On the other hand, a much publicized programme
like the Bolsa Família amounts to only a tiny fraction of the GDP.

TABLE 2
Selected Government Expenditures (% of GDP) – Brazil, 2006
Expenditures Share of GDP (%)
Public education 3.8
Social Security and pensions 11.1
Private sector 6.8
Civil servants 4.3
Social Assistance cash transfers 0.8
Benefício de Prestação Continuada (BPC) 0.4
Programa Bolsa Família (PBF) 0.4
Total 15.7
Total tax revenue 34.1
Source: Mostafa, J; Souza, PHGF; Vaz, FM. Efeitos econômicos do gasto social. In: Castro, JA; Ferreira, H; Campos, AG;
Ribeiro, JAC (Org). Perspectivas da Política Social no Brasil. Brasília: Ipea, 2010. Total tax revenue from Ribeiro, MB.
Uma análise da carga tributária bruta e das transferências de assistência e previdência no Brasil no período 1995-2009:
evolução, composição e suas relações com a regressividade e a distribuição de renda. In: Castro, JA; Santos, CHM;
Ribeiro, JAC. Tributação e eqüidade no Brasil: um registro da reflexão do Ipea no biênio 2008-2009.Brasília: Ipea, 2010.

EDUCATION
Historically, Brazil has been plagued by overall low levels of educational attainment and a very
unequal distribution of educational opportunities favouring the upper-middle class and the
rich. Most of the latter half of the 20th century was characterized by very limited access to
primary education, especially in the north and northeast regions, while the government

Working Paper 9

poured substantial resources on free public post-secondary education. It is no wonder, then,
that education has been singled out for decades as the main determinant of inequality in Brazil
(Langoni, 1973; Jallade, 1978; Ramos, 1991; Barros; Mendonça, 1995; Ferreira, 2000; Barbosa
Filho & Pessoa, 2008).
The Brazilian educational system has undergone extensive reforms since adoption of the
1988 Constitution. A stronger focus on primary and secondary education was accompanied by
more effective funding arrangements.5 Public expenditures on education added up to only 2.7
per cent of the GDP in 1980, but, since the mid-1990s, they have hovered around 4 per cent or
4.5 per cent. One of the results was that school attendance for children between 7 and 14 years
old—which is mandatory—finally became nearly universal by the late 1990s.
Amid the recent economic growth, there was not only an increase in real terms in public
investment on education per student, but such investment also became considerably less
biased toward post-secondary education. According to data from the Ministry of Education,
the real public investment per student expanded by 94 per cent between 2000 and 2009—
almost entirely after 2003/2004—while the ratio between expenditures on post-secondary and
primary students dropped from a lofty 11.1 times to a more palatable 5.2 times.6
All these developments were reflected on quickly improving educational standards, as
seen in Tables 3 and 4. Illiteracy rates remain high, mostly due to the inertial influence of past
generations; among the 15-to-24 age bracket, it is already quite low. Attendance rates
improved across the board, especially among teenagers between 15 and 17. This has already
resulted in a much more educated labour force: only about one third of the economically
active population in 1995– which comprises those who are employed and the unemployed
who are actively looking for jobs—had completed mandatory primary education; that per
centage almost doubled in less than 15 years. Similarly, the supply of workers with higher
educational credentials also enlarged significantly.

TABLE 3
Illiteracy and Attendance Rates (%) – Brazil, 1995 and 2009
1995 2009
Illiteracy rates
Ages 15+ 15.5 9.7
Ages 15‐24 7.1 1.9
Attendance rates
Ages 6‐14 88.7 97.6
Ages 15‐17 66.7 85.2

TABLE 4
Educational Attainment of the Economically Active Population (%) – Brazil, 1995 and 2009
Completed at least… 1995 2009
Primary Education 34.5 61.7
Secondary Education 20.7 44.1
Tertiary Education 5.6 10.2


The pace of the changes can be better appreciated in Figure 5. In less than 15 years, the
mean number of years of schooling7 increased from 5.8 to 8.3, a 42 per cent hike. Unlike in the
1960s, when there also an expansion of the schooling system, this time the heterogeneity in
educational attainment diminished: the Gini index of years of schooling fell abruptly by 30
per cent, from 0.413 in 1995 to 0.288 in 2009. In other words, for the first time in decades, Brazil
finally succeeded in boosting educational levels and in providing less unequally distributed
educational opportunities. This has been one of the pivotal reasons for the rapid decrease of
labour market inequality in the past few years and has also contributed significantly to the
overall reduction in household per capita income inequality (Barros, Franco & Mendonça, 2006).

FIGURE 5
Mean years of Schooling and Gini Index of Years of Schooling for the
Economically Active Population – Brazil, 1995/2009

Source: Pesquisa Nacional por Amostra de Domicílios, 1995/2009.

Even better, as this is a long-term policy that has already spanned several presidential
mandates, there is reason to believe that this trend will probably persist in the foreseeable
future. This is absolutely crucial for Brazil’s development because there is still much to be done.
On the one hand, 8.3 mean years of schooling is still very low, as it barely covers mandatory
primary education (8 years). On the other hand, in spite of recent improvements, Brazilian
students’ scores on standardized international cognitive tests are still lagging behind, which
reflects the poor quality of the public schooling system (Soares & Nascimento, 2011).

MINIMUM WAGE LAW8
The federal Minimum Wage Law was first enacted in Brazil in the late-1930s and established
more than a dozen regional minimum wages in order to account for regional differences in
cost of living. Initially, it applied only to formal urban workers, being extended to formal rural
workers— only a tiny minority of the massive rural workforce—only in the 1960s. Later, in
1984, a new unified national minimum wage was established.9

Working Paper 11

For decades, there were no laws regulating minimum wage adjustments, which
were granted at the president’s discretion. As such, its real value has fluctuated wildly.
Many authors who have studied the increase in income inequality in the 1960s, for instance,
blame specifically the military government’s decision to depreciate the minimum wage in real
terms during that decade (Fishlow, 1972; Hoffmann & Duarte, 1972).
Since adoption of the 1988 Constitution, the importance of the minimum wage grew
considerably. On the one hand, the Constitution reaffirmed the existence of a single national
minimum wage, applicable to all (formal) workers, and stipulated that it should be periodically
adjusted in order to recoup losses due to inflation. On the other hand, the minimum wage also
became tied to other government programmes, being enshrined as the minimum value of all
Social Security and some Social Assistance benefits.
During the high inflation period of the 1980s, the minimum wage was adjusted monthly
to try to recoup losses, with no real gains. After the macro-economic stabilization, its real value
started to increase consistently, albeit slowly. Since the mid-2000s, with the economic recovery
in full swing, the government and employers’ and workers’ associations finally came to an
agreement that was set to provide rapid real gains: each annual minimum wage adjustment
would recoup the previous year’s inflation, compounded with the real GDP growth of the year
before the last.10
This agreement, which was signed into law earlier in 2011, has resulted in larger real
minimum wage gains over the past few years (see Figure 6). The period between the unification
of the national minimum wage and the macro-economic stabilization had—if anything—a
declining real minimum wage. Afterwards, between 1995 and 2005, the real value of the
minimum wage grew at a rate of roughly 7 per cent per year. Finally, between 2005 and 2011,
this rate rose to 10 per cent per year. Overall, the real minimum wage in early 2011 was almost
three times higher than in 1995, which implies an annual growth rate of about 8 per cent.
This has also contributed to the decline in labour market inequality and overall
income inequality and poverty (Foguel et al., 2000; Firpo; Reis, 2006; Saboia, 2007; see below).
In fact, the minimum wage nowadays is high enough to ensure that only very few families with
at least one recipient remain extremely poor. Only very large families—which are increasingly
rare—with at least one formal worker or one beneficiary of Social Security pensions remain
below the poverty line.
According to the PNAD, there were about 9 million workers who received the minimum
wage as remuneration in 2009, adding up to roughly 11 per cent of the employed labour force.
Furthermore, more than 13 million individuals—almost 60 per cent of all pensioners—had
retirement benefits equal to the minimum wage and another 1.5 million people benefitted from
the Benefício de Prestação Continuada (BPC), a Social Assistance benefit paid to poor people over 65
or with a disability that is also tied to the minimum wage.11 These benefits are heavily subsidized
by the federal government and profoundly redistributive, though expensive (see Table 1).
More recently, it has been speculated that the minimum wage policy has reached its
limits for two main reasons: new adjustments imply a mounting fiscal burden, while their
positive redistributive effects are rapidly diminishing because the poorest groups remain
disconnected from the formal labour market and Social Security (Barros, 2007; Giambiagi &
Franco, 2007). There is certainly a grain of truth to those accounts, but, so far, there are no
plans to modify the current policy.


FIGURE 6
Real Monthly Minimum Wage (US$ PPP) – Brazil, 1985/2011

* The shaded area marks the post-stabilization period.

SOCIAL SECURITY
Brazilian Social Security dates back to the late-19th century industry-specific Funds for
Retirement and Pensions, which were progressively unified under a framework inspired by
the Bismarckian German model. It became fully state-run in the 1960s and entirely separate
from the health care system only after adoption of the 1988 Constitution.12
To this day, it has at least two main branches:—one for private sector workers and one
for civil servants. As a mandatory and contributory ‘pay-as-you-go’ system that benefits mostly
formal workers, it has traditionally left out a considerable proportion of the Brazilian population.
Since adoption of the 1988 Constitution, however, the Private Sector branch has been
expanded considerably. For instance, the so-called ‘Rural Social Security’, which is almost non-
contributory, as it encompasses mostly small farmers and poor rural workers, went from 4 million
monthly benefits in 1991 to 7 million in 2003, a 75 per cent increase in just 12 years (Ipea, 2009).
This helped to reduce income inequality and poverty in rural areas. Likewise, tying its minimum
benefits to the minimum wage was also crucial (see above). More recently, the rapid creation
of formal jobs has been another key factor in enlarging the reach of the Social Security.
All of these reforms turned the Brazilian Social Security, or at least its Private Sector
branch, into a powerful tool to reduce poverty among the elderly. In 2009, about 90 per cent
of the population over 65 received a Social Security benefit and poverty levels were below
one per cent for this group. In contrast, among children 15 or younger the poverty levels
were much higher, around 8 per cent.

Working Paper 13

According to the PNADs, about 14 million individuals—slightly under 10 per cent of the
population—benefitted from Social Security pensions in September of 1995; in 2009, there
were already over 22 million recipients, or about 12 per cent of the population. In absolute
terms, this represents a 59 per cent expansion of the number of pensioners. The average
pension value also climbed sharply, from US$ PPP 305 in 1995 to US$ PPP 575 in 2009, an 88
per cent raise. Hence, Social Security pensions went from 13 per cent of the total income
reported in the PNADs in 1995 to almost 19 per cent in 2009.
The flipside of this system is that it amounts to a disproportionate per centage of the GDP
when compared to other countries and, worse, runs significant annual deficits: about 1.3 per cent
of GDP for the Private Sector and 2.1 per cent for the Civil Servants’ Social Security (Ipea, 2011).
This and the general ageing of the population have put the Social Security system under
scrutiny, with recent reforms trying to limit expenses by tightening the retirement conditions.
It can be argued that the deficits are not a particularly worrisome issue for the Private
Sector Social Security, as those can be partially swayed if the recent trend of formalization
continues. Also, the benefits paid are generally progressive and very important for alleviating
poverty among the elderly.
On the other hand, the Civil Servants’ Social Security covers just a tiny fraction of the
population and its large pay-cheques actually worsen income inequality (Silveira, 2008;
Rangel, 2011). Therefore, those deficits are far more troublesome. It is still too early to assess
the impact of the 2003 reform, but preliminary evaluations suggest it may have far-reaching
positive consequences (Soares, 2011).

SOCIAL ASSISTANCE
Traditionally, social assistance programmes in Brazil have been highly fragmented and
spearheaded by non-profit charitable foundations, with government interventions either
restricted to particular emergencies, like droughts, or marred by clearly clientelist practices.
As with other areas, this started to change after adoption of the 1988 Constitution, which
enshrined Social Assistance as one of the fundamental social rights.
To this day, Social Assistance consists of two major areas: the provision of services
and targeted cash transfers. There is not much to say about the former, as it is still—quite
understandably—largely a work in progress, considering that the Sistema Único de Assistência
Social (SUAS – Unified System of Social Assistance) dates back just to 2004 and its legal
framework is still being laid out.13
Cash transfer programmes, in contrast, are in a certain sense much more developed.
Due to their widespread popularity, they have been the most visible side of social assistance in
Brazil since the mid-1990s. There are two major federal programmes: the Benefício de Prestação
Continuada (BPC – Continuous Cash Benefit) and the Programa Bolsa Família
(Family Benefit Programme).
The BPC was the earliest one, created by the 1988 Constitution to replace its more modest
predecessor, the Renda Mensal Vitalícia (RMV – Lifelong Monthly Income). Legally, though, it
was fully spelled out only in 1993 and began its transfers only in 1996, paying benefits equal to
the minimum wage.


As an unconditional monthly cash transfer, the BPC is targeted to those with family per
capita income that is one quarter below the minimum wage and who belong to one of two
groups: individuals of any age with severe disabilities or the elderly over 65. Because the
transfer is a constitutional right, anyone who meets the eligibility criteria must automatically
receive the benefit after applying for the programme. One the main advantages of this legal
status is that it has assured significant political independence for the programme (Medeiros,
Diniz & Squinca, 2006; Medeiros, Britto & Soares, 2008).
The Programa Bolsa Família is by far the most renowned cash transfer programme in
Brazil. Its origins date back to municipal-level experiences in the mid-1990s, which were then
espoused by the federal government in the late-1990s as a series of fragmented programmes
that were finally unified under the aegis of the Programa Bolsa Família in late-2003. Unlike the
BPC, the Bolsa Família is not an entitlement: the number of beneficiaries depends largely on
budget constraints and thus it is possible that an eligible family may apply but be denied the
benefit (see Soares et al., 2006; Soares et al., 2007; Soares, 2011) .
The Bolsa Família has two types of benefits that are targeted to somewhat different groups.
As of 2010, extremely poor families—families whose per capita income was under approximately
US$ PPP 40 per month—were eligible to receive an unconditional basic transfer of US$ PPP
38.40. In addition, all families with children of up to 17 years old whose income per capita was no
higher than US$ PPP 80 were eligible to receive US$ PPP 12.40 per child up to 15 years old (with a
maximum of three benefits per family)14 and US$ PPP 18.65 per youngster between 16 and 17
years old (with a maximum of two benefits per family). The maximum benefit value—for
extremely poor families with at least three children younger than 16 and another two younger
than 18—was therefore US$ PPP 113 per family.15 The conditionalities for the children’s benefits
deal mostly with school attendance, children’s immunizations and pre- and post-natal care.
Tables 5 and 6 present information on the BPC and Bolsa Família for 2010. Due to the real
gains in the minimum wage, the BPC’s budget became considerably greater than that of the
Bolsa Família (compare with Table 2 above).
Each programme is based on radically different reasoning: focused on individuals, the
BPC has a much smaller number of recipients, who, in turn, receive a much larger monthly sum.
The Bolsa Família has a much larger clientele, but its benefits pale in comparison. Whereas a
beneficiary of the BPC is almost certain to be above the income poverty line, the same is not
true of Bolsa Família families: for a typical four-person family, the mean benefit amounts to less
than US$ PPP 14 per capita, which is only 35 per cent of the poverty line. In order words, while
the BPC alone can almost always lift its recipients’ families out of poverty, the Bolsa Família
families can escape extreme poverty only if they have income from other sources.

TABLE 5
Benefício de Prestação Continuada (BPC) – Brazil, 2010
Target # of benefits 2010 budget Eligibility line Mean monthly benefit
Group (December) (% of GDP) (family per capita income, US$ PPP) (per individual, US$ PPP)
Elderly 1.8m 0.28
Disability 1.6m 0.26 72 288
Total 3.4m 0.55
Source: Ministry of Social Development.

Working Paper 15

TABLE 6
Programa Bolsa Família – Brazil, 2010
# of families 2010 budget Eligibility line Mean monthly benefit
Programme
(December) (% of GDP) (family per capita income, US$ PPP) (per family, US$ PPP)
Bolsa 12.8m 0.39 40 (no children) 55
Família 80 (with children)
Source: Ministry of Social Development.

The BPC and especially the Bolsa Família have been extensively studied. Both seem to be
extraordinarily well-targeted, insofar as the inclusion errors are only moderate and mainly
involve the ‘quasi-poor’. Exclusion errors, in contrast, seem to be larger, which suggests that
both programmes could be scaled up. Negative effects on the incentives to work and Social
Security contributions have been found to be either non-existent or otherwise irrelevant.
The same applies to fertility rates: Bolsa Família has apparently no effect on them. Finally,
both programmes are tremendously pro-poor and have helped decisively to reduce income
inequality and poverty (Barros, Carvalho & Franco, 2006; Soares & Sátyro, 2009; Soares et al.,
2009; Soares et al., 2010).
While this last point will be more extensively discussed in the next section, it is very well-
illustrated by Figure 7.16 The BPC and the Bolsa Família are stunningly well-targeted. This is no
mean feat, especially for the latter, as its sheer size presumably should make its accuracy quite
hard to maintain. As mentioned, most of the supposedly non-eligible beneficiaries are
relatively close to the eligibility lines and thus it is plausible that—due to income volatility—
they were in fact eligible at some point in time.

FIGURE 7
Individuals who Benefit directly or Indirectly from Transfers by Centiles of Household
Per Capita Income – Brazil, 2009

* The shaded area marks the upper eligibility line of the Programa Bolsa Família (US$ PPP 80).


6 DECOMPOSING INCOME INEQUALITY
One general way of evaluating the impact of those four kinds of interventions—education,
minimum wage, Social Security and Social Assistance—on income inequality is through the
standard decompositions of the Gini and Generalized Entropy indices. Although somewhat
crude, they are simple and very informative and therefore remain one of the most-used tools
of Brazilian researchers.

FORMULAS
In the first case, the Shorrocks (1982) decomposition of the Gini index by factor components
allows us to assess the relative contribution of different types of income to the overall
inequality and to changes in inequality over time.
For the static decomposition—the decomposition of the Gini index at one point in time—
the formula is quite straightforward. Suppose that an individual’s income per capita ( x i ) is
k
equal to the sum of all incomes he or she receives from h different sources ( xi = ∑ x hi ); then
h =1
the Gini index can be expressed as the sum of each income source’s coefficient of
μh
concentration weighted by its relative income share ( φ h = ):
μ

k
G = ∑ φh Ch
h =1 (1)

The coefficient of concentration of each income component, which measures how
unequally distributed it is, varies from -1 (all income from that factor component goes to the
overall poorest person in the population) to +1 (all income from that factor component goes to
the richest person). If one income source has a perfectly egalitarian distribution, its coefficient
of concentration will be zero. As a rule of thumb, any income component having a coefficient
of concentration higher than the overall Gini index can be considered regressive—i.e., it
increases income inequality—and vice-versa. The formula for the coefficient is:

2 x
Ch = cov(i, hi )
n μh (2)

Of course, this formula is very similar to the standard algebraic formula of the Gini index itself:

2
G= cov(i, xi )
nμ (3)

Working Paper 17

The dynamic decomposition compares changes in the Gini index between two points in
time and can be expressed as:17

k
ΔG = ∑ ((C h − G ) Δφ h + ΔC hφ h )
h =1 (4)

The first term captures, for each h income component, the ‘composition effect’, that is, the
change in the Gini index caused by variations in the relative income share of that component.
The second term is the ‘concentration effect’, that is, the change in the Gini index resulting
from variations in the distribution of that component.
The second type of inequality decomposition that is also very useful is the
decomposition by population subgroups of the Generalized Entropy family of inequality
indices (see Mookherjee & Shorrocks, 1982). Given the prominence of regional, racial, and
educational disparities in Brazil, this type of decomposition has always been interesting.
In this paper, the decomposition will be applied to the GE(0) index—also known as the
‘Theil-L index’ or the ‘mean log deviation’—because it allows a counterfactual interpretation
of between-group inequality. In other words, if the inequality between two groups, A and B, is
responsible for X per cent of total inequality, then the total inequality would be X per cent
lower if the average income of both groups were the same.
The static and dynamic decompositions of the GE(0) index are given by:

k k
1
GE (0) = I within − group + I between− group = ∑ f j GE (0) j + ∑ f j ln( )
j =1 j =1 λj
(5)

k k k k
ΔGE(0) = ∑ f j ΔGE(0) j + ∑ Δf j GE(0) j + ∑ Δf j [λ j − ln(λ j )] + ∑(v j − f j )Δ ln(μ j )
j =1 j =1 j =1 j =1
(6)

Where f j is the population share of the subgroup j; GE (0) j is the GE(0) index for
μj
subgroup j; λ j is the relative mean income of subgroup j ( ); v j is the income share of
μ
subgroup j; and μ j is the mean income of subgroup j.

In the dynamic decomposition, the first term is the ‘pure inequality’ effect, that is, the
contribution of changes in intra-subgroup inequality to the overall change in inequality;
the second and third terms are the ‘allocation’ effect, that is, the effect caused by changes
in the relative sizes of the subgroups; and the fourth term is the ‘income’ effect, that is, the result
of changes in the subgroups’ relative mean incomes.


GINI DECOMPOSITION BY FACTOR COMPONENTS
Because income inequality started to fall around 2001 (see Figure 4), the inequality decomposition
presented in this section will compare that year with 2009. The relevant income variable is the
household income per capita, which can be divided into seven components: labour income
equal to the minimum wage; other labour income; pension income equal to the minimum
wage; other pension income:18 Bolsa Família (and its predecessors) income; BPC income;
and other incomes.
Table 7 shows the results of the static decompositions for 2001 and 2009. Starting with
the coefficients of concentration, it is clear that all income components tied to the minimum
wage are very progressive—i.e., pro-poor, helping to decrease income inequality—although
the real minimum wage gains over the decade have made them slightly less so. The Bolsa
Família transfers also stand out: they were already extremely progressive in 2001 and became
even more so during the decade.
In contrast, the Social Security pensions higher than the minimum wage have been found
to be the most regressive income component of all seven. This is due mostly to the Public
Sector (Civil Servants’) Social Security and exacerbates inequality. This is particularly startling,
as Social Security in Brazil is a state-subsided ‘pay-as-you-go’ system without individual
accounts, so that a worker’s contribution to Social Security bears only an imperfect relation to
his or her future benefits. In this sense, the high, positive coefficient of concentration suggests
that parts of the state’s expenditures are actually helping to increase income inequality.
Furthermore, all regressive income sources—labour and pension income higher than
minimum wage and other income—have remained stable or become less unequally
distributed. For example, the coefficient of concentration for labour income higher than
the minimum wage dropped from 0.608 to 0.576 and thus became more pro-poor (or less
regressive). This stems mainly from the educational improvements mentioned above.

TABLE 7
Gini Decomposition by Factor Components of Household
Per Capita Income – Brazil, 2001 and 2009
Coefficient of
Income share Contribution to Gini % of Gini
Income sources concentration
2001 2009 2001 2009 2001 2009 2001 2009
Minimum ‐0.115 ‐0.091 0.021 0.036 ‐0.002 ‐0.003 ‐0.4 ‐0.6
wage
Labour
Other 0.608 0.576 0.759 0.726 0.461 0.418 77.7 77.5

Minimum 0.097 0.157 0.037 0.057 0.004 0.009 0.6 1.7
wage
Pensions
Other 0.743 0.742 0.134 0.131 0.099 0.097 16.7 18.0

Bolsa Família and ‐0.315 ‐0.526 0.001 0.007 0.000 ‐0.004 0.0 ‐0.7
predecessors

BPC ‐0.081 ‐0.016 0.001 0.006 0.000 0.000 0.0 0.0

Other 0.672 0.603 0.048 0.037 0.032 0.022 5.4 4.2

Gini 1.000 1.000 0.594 0.539 100 100

Working Paper 19

In view of each component’s income share, the most salient change over the decade was
that all pro-poor income sources increased their participation in the total income reported by
the PNADs. Thus, the decline in income inequality observed between 2001 and 2009 was made
possible not only by a reduction in the coefficient of concentration of the regressive transfers,
but also by an expansion of the income share of the inequality-reducing components.
Table 8 features the dynamic decomposition of the changes in the Gini index between
2001 and 2009. Overall, changes in labour income were responsible for 63 per cent of the
decline in income inequality, highlighting the pivotal role of the educational and the minimum
wage policies. The BPC and the Bolsa Família added up to another 18 per cent, which is no
mean feat, in view of the fact that they stand for just slightly over 1 per cent of the total
reported income. Finally, Social Security, the one social intervention with dubious results
in the static decomposition, also played a relevant role, contributing to 11 per cent of the
fall in inequality, thanks mostly to the expansion of benefits indexed to the minimum wage.

TABLE 8
Dynamic Decomposition of the Gini Index of Household
Per Capita Income – Brazil, 2001-2009
Composition Concentration
Income sources Total As % of ∆Gini
effect effect

Minimum wage ‐0.010   0.001   ‐0.010   17.9
Labour
Other ‐0.001   ‐0.024   ‐0.025   45.5

Minimum wage ‐0.009   0.003   ‐0.006   10.5
Pensions
Other 0.000   0.000   ‐0.001   1.0

Bolsa Família and predecessors ‐0.006   ‐0.001   ‐0.007   12.7

BPC ‐0.003 0.000   ‐0.003   5.7

Other ‐0.001   ‐0.003   ‐0.004   6.7

Total ‐0.031 ‐0.024 ‐0.055 100


In short, it is very encouraging to see Brazil’s major social programmes getting more
effective when it comes to reducing income inequality. The real challenge now, of course, is to
remain on track, as it might get increasingly more difficult to keep this pace. Again, one reason
for optimism is that there is probably still room for further decline in labour market inequality as
long as the current educational and minimum wage policies are maintained and improved upon.

DECOMPOSITION BY POPULATION SUBGROUPS
It is appropriate to further investigate the decisive role that labour market inequality plays in
overall income inequality. The method of choice is the decomposition of the GE(0) index of
labour income inequality among all employed workers with positive earnings. Unlike in the
previous section, the comparison is now between 2002 and 2009.19


Six of the most-studied inequality-generating dimensions were considered: schooling
(16 population subgroups, from 0 to 15 years of schooling); industry (eight sectors: agriculture;
manufacturing; construction; services; education, health and social services; domestic services;
government; other); race (five categories: white, black, brown, Asian, indigenous); regional
segmentation (27 categories: 26 states and the federal district); urban/rural areas; and gender
(male/female). Moreover, a seventh dimension was generated by combining three of the
most important dimensions (schooling, industry and states), which partitioned the population
into 3,456 subgroups.
The results for the static GE(0) decomposition are presented in Table 9. Between 2002
and 2009, the GE(0) index of inequality for all employed workers with positive earnings fell
by 16 per cent. Likewise, the absolute value for all between-group components also declined.
In most cases, it actually fell faster than overall inequality, and therefore the per centage of the
total inequality accounted for by between-group differences decreased over time. In 2002, for
instance, the earnings inequality between workers with varying years of schooling accounted
for 47 per cent of the total earnings inequality; in 2009, this per centage had dropped to
43 per cent, as the between-group inequality declined faster than the intra-group inequality.
This is a very consequential development because it suggests that the structure of
earnings inequality is possibly being reshaped. This is especially true regarding educational
credentials, which are shown to be far more relevant than the other dimensions.
Finally, the fact that the explanatory power of the combined schooling/industry/state
partition is much lower than the sum of each dimension’s between-group component should
serve as a reminder of the strong correlation among several of the major dimensions of
inequality in Brazil.

TABLE 9
GE(0) Decomposition Labour Income by Population Subgroups for all Employed Workers with
Positive Earnings – Brazil, 2002 and 2009
Absolute Relative

2002 2009 ∆ (%) 2002 2009 ∆ (p.p.)

GE(0) 0.582 0.491 ‐16 100 100 ‐

Between‐Group Components

Schooling + Industry   0.275   0.211   ‐23   47.2   43.1   ‐4.1
+ State

Years of schooling   0.209   0.158   ‐24   35.9   32.2   ‐3.9
(16 groups)

Industry (8) 0.085   0.072   ‐15   14.6   14.8   +0.2

Race (5) 0.057   0.039   ‐33   9.9   7.9   ‐2.0

State (27) 0.050   0.032   ‐36   8.6   6.6   ‐2.0

Urban/rural areas 0.033   0.019   ‐43   5.6   3.8   ‐1.8

Male/female 0.014   0.013   ‐5   2.4   2.7   +0.3


Working Paper 21

The dynamic decomposition featured in Table 10 shows how changes in the educational
attainment of the workforce contributed to the decline in earnings inequality. The educational
improvement documented above had a negative allocation effect, but a more homogeneously
educated labour force enjoyed a dominant income effect, as declining returns to education
narrowed the income gaps among the different levels of educational attainment. Within-group
inequality—which, in this case, can be best interpreted as a residual—also contributed to the
overall drop of the GE(0) index.

TABLE 10
Dynamic Decomposition of the GE(0) Index of Labour Income by Educational Subgroups –
Brazil, 2002-2009
∆2002‐2009 %
Pure inequality effect ‐0.041 45.3
Allocation effect 0.013 ‐14.9
Income effect ‐0.062 69.2
Total ‐0.091 100

7 CONCLUSIONS
Since the mid-1990s, and especially in the past decade, Brazil has undergone a period of rapid
poverty and inequality reduction that was made possible by a consumer-led economic boom
as well as more effective social policies. Nonetheless, Brazil is still a middle-income country
with an unacceptably high level of income inequality, and thus it is imperative to remain on
track and keep the recent trajectory of pro-poor growth going.
As with all countries, this means that the challenges will probably be increasingly
difficult. But there are reasons for optimism. Social expenditures are one of them: the renewed
commitment to social programmes since adoption of the 1988 Constitution has largely turned
them into extremely valuable tools to reduce poverty and inequality, although they do come
at a hefty price. Still, educational and minimum wage policies have greatly benefitted the
labour market, while Social Security and Social Assistance form a social protection network
that has greatly diminished poverty among the elderly and, to a lesser extent, children.
Consequently, these policies and programmes have generally played a central role in
bringing about the reduction in income inequality during the past decade.
Furthermore, it is clear that the social programmes as a whole still have plenty of room
for improvement. As argued above, Bolsa Família is formidable, but its benefits are too low
and it could be scaled up, as some eligible families are still excluded from the programme.
Similarly, educational attainment is still very low and the overall quality of public schools is
substandard. A lot could be done in both areas if, for instance, Civil Servants’ Social Security
were fixed, as it is inordinately expensive and runs huge annual deficits that drain resources
away from pivotal areas. Finally, some policies that could a lot to further reduce poverty and
inequality—such as land reform—have largely been cast aside, so it would do wonders
to bring them back to the political agenda.


REFERENCES
Abreu, M. De P. The Brazilian Economy, 1928-1980. Rio De Janeiro: Puc-Rio, 2000.
(Texto Para Discussão, N. 433)
Baer, W. A Economia Brasileira. São Paulo: Nobel, 2009.
Barbosa Filho, F. De H.; Pessoa, S. Retorno Da Educação No Brasil. Pesquisa E Planejamento
Econômico, V. 38, N. 1, P. 97-125, 2008.
Barros, R. P. A Efetividade Do Salário Mínimo Em Comparação À Do Programa Bolsa Família
Como Instrumento De Redução Da Pobreza E Da Desigualdade. In: Barros, R. P. Foguel, M. N.;
Ulyssea, G. (Eds.). Desigualdade De Renda No Brasil: Uma Análise Da Queda Recente.
Brasília: Ipea, 2007. V. 2.
Barros, R. P. Carvalho, M.; Franco, S. O Papel Das Transferências Públicas Na Queda Recente Da
Desigualdade De Renda Brasileira. In: Barros, R. P. Foguel, M. N.; Ulyssea, G. (Eds.). Desigualdade
De Renda No Brasil: Uma Análise Da Queda Recente. Brasília: Ipea, 2007. V. 2.
Barros, R. P. Et Al. A Queda Recente Da Desigualdade De Renda No Brasil. In: Barros, R. P. Foguel,
M. N.; Ulyssea, G. (Eds.). Desigualdade De Renda No Brasil: Uma Análise Da Queda Recente.
Barros, R. P. Cury, S.; Ulyssea, G. A Desigualdade De Renda No Brasil Encontra-Se Subestimada?
Uma Análise Comparativa Usando Pnad, Pof E Contas Nacionais. In: Barros, R. P. Foguel, M. N.;
Ulyssea, G. (Eds.). Desigualdade De Renda No Brasil: Uma Análise Da Queda Recente.
Barros, R. P. Franco, S.; Mendonça, R. A Recente Queda Na Desigualdade De Renda E O
Acelerado Progresso Educacional Brasileiro Da Última Década. In: Desigualdade De Renda No
Brasil: Uma Análise Da Queda Recente. Brasília: Ipea, 2007. V. 2.
Barros, R. P.; Mendonça, R. Os Determinantes Da Desigualdade No Brasil.
Rio De Janeiro: Ipea, 1995. (Texto Para Discussão, N. 377).
Dieese. Salário Mínimo: Instrumento De Combate À Desigualdade. São Paulo: Dieese, 2010.
Ferreira, F. H. G. Os Determinantes Da Desigualdade De Renda No Brasil: Luta De Classes Ou
Heterogeneidade Educacional? In: Henriques, R. (Ed.). Desigualdade E Pobreza No Brasil.
Rio De Janeiro: Ipea, 2000. .
Firpo, S.; Reis, M. O Salário Mínimo E A Queda Recente Da Desigualdade No Brasil. In: Barros, R.
P. Foguel, M. N.; Ulyssea, G. (Eds.). Desigualdade De Renda No Brasil: Uma Análise Da Queda
Recente. Brasília: Ipea, 2006. V. 2.
Fishlow, A. Brazilian Size Distribution Of Income. The American Economic Review, V. 62, N. 1-2, P.
391-402, 1972.
Fishlow, A. Brazilian Development In Long-Term Perspective. The American Economic Review,
V. 70, N. 2, P. 102-108, 1980.
Foguel, M. N. Et Al. Uma Avaliação Dos Impactos Do Salário Mínimo Sobre O Nível De Pobreza
Metropolitana No Brasil. Rio De Janeiro: Ipea, 2000. (Texto Para Discussão, N. 739).

Working Paper 23

Giambiagi, F.; Franco, S. O Esgotamento Do Papel Do Salário Mínimo Como Mecanismo De
Combate À Pobreza Extrema. Rio De Janeiro: Ipea, 2007. (Texto Para Discussão, N. 1290).
Hoffmann, R.; Duarte, J. C. A Distribuição Da Renda No Brasil. Revista De Administração De
Empresas, V. 12, N. 2, P. 46-66, 1972.
Ipea. Políticas Sociais: Acompanhamento E Análise - Vinte Anos Da Constituição Federal.
Brasília: Ipea, 2009.
Ipea. Políticas Sociais: Acompanhamento E Análise, N. 19. Brasília: Ipea, 2011.
Jaccoud, L. Hadjab, P. D. E.-M.; Chaibub, J. R. The Consolidation Of Social Assistance In Brazil And
Its Challenges, 1988-2008. Brasília: International Poverty Centre For Inclusive Growth, 2010.
(Working Paper, N. 76).
Jallade, J.-P. Educação De Primeiro Grau E Desigualdade De Renda No Brasil: Perspectivas A
Longo Prazo. Pesquisa E Planejamento Econômico, V. 8, N. 3, P. 561-598, 1978.
Langoni, C. Distribuic Ão Da Renda E Desenvolvimento Econo Mico Do Brasil.
Rio De Janeiro: Editora Expressão E Cultura, 1973.
López-Calva, L. F.; Lustig, N. Explaining The Decline In Inequality In Latin America:
Technological Change, Educational Upgrading, And Democracy. In: López-Calva, L. F.; Lustig, N.
(Eds.). Declining Inequality In Latin America: A Decade Of Progress? Baltimore, Eua: Brookings
Institution Press, United Nations Development Programme, 2010.
Martine, G. O Papel Dos Organismos Internacionais Na Evolução Dos Estudos Populacionais No
Brasil: Notas Preliminares. Revista Brasileira De Estudos De População, V. 22, N. 2, P. 257-275, 2005.
Martine, G. Et Al. A Pnad: Notas Para Uma Avaliação. In: Sawyer, D. (Ed.). Pnads Em Foco: Anos 80.
Belo Horizonte: Abep, 1988. P. 281-305.
Medeiros, M. Britto, T.; Soares, F. V. Targeted Cash Transfer Programmes In Brazil: Bpc And The Bolsa
Família. Brasília: International Poverty Centre For Inclusive Growth, 2008. (Working Paper, N. 46).
Medeiros, M. Diniz, D.; Squinca, F. Cash Benefits To Disabled Persons In Brazil: An Analysis Of The
Bpc - Continuous Cash Benefit Programme. Brasília: International Poverty Centre For Inclusive
Growth, 2006. (Working Paper, N. 16).
Médici, A. A Mensuração Da Subjetividade: Notas Sobre A Variável Renda Nas Pnads. In: Sawyer,
D. (Ed.). Pnads Em Foco: Anos 80. Belo Horizonte: Abep, 1988. P. 121-151.
Milanovic, B.; Yitzhaki, S. Decomposing World Income Distribution: Does The World Have A
Middle Class? The Review Of Income And Wealth, V. 48, N. 2, P. 155-178, 2002.
Mookherjee, D.; Shorrocks, A. A Decomposition Analysis Of The Trend In Uk Income Inequality.
The Economic Journal, V. 92, N. 368, P. 886-902, 1982.
Ramos, L. Educação, Desigualdade De Renda E Ciclo Econômico No Brasil. Pesquisa E
Planejamento Econômico, V. 21, N. 3, P. 423-448, 1991.
Rangel, L. A. Aspectos Distributivos Do Regime De Previdência Dos Servidores Públicos.
Rio De Janeiro: Ipea, 2011. (Texto Para Discussão, N. 1617).
Rocha, R. De R.; Caetano, M. A.-R. O Sistema Previdenciário Brasileiro: Uma Avaliação De
Desempenho Comparada. In: Caetano, M. A.-R. (Ed.). Previdência Social No Brasil: Debates E
Desafios. Brasília: Ipea, 2008.


Rocha, S. A Investigação Da Renda Nas Pesquisas Domiciliares. Economia E Sociedade, V. 12,
N. 2, P. 205-224, 2003.
Saboia, J. Efeitos Do Salário Mínimo Sobre A Distribuição De Renda No Brasil No Período
1995-2005 - Resultados De Simulações. Econômica, V. 9, N. 2, P. 270-295, 2007.
Sayad, J. Observações Sobre O Plano Real. Estudos Econômicos, V. 25, N. Especial,
P. 7-24, 1996. 1995.
Shorrocks, A. Inequality Decomposition By Factor Components. Econometrica, V. 50, N. 1,
P. 193-211, 1982.
Silveira, F. G. Tributação, Previdência E Assistência Sociais: Impactos Distributivos. Campinas:
Universidade De Campinas, 2008.
Soares, F. V. Bolsa Família: A Review. Economic & Political Weekly, V. Xlvi, N. 21, P. 55-60, 2011.
Soares, F. V. Et Al. Cash Transfer Programmes In Brazil: Impacts On Inequality And Poverty.
Brasília: International Poverty Centre For Inclusive Growth, 2006. (Working Paper, N. 21).
Soares, F. V. Et Al. Programas De Transferência De Renda No Brasil: Impactos Sobre A
Desigualdade. In: Barros, R. P. Foguel, M. N.; Ulyssea, G. (Eds.). Desigualdade De Renda No Brasil:
Uma Análise Da Queda Recente. Brasília: Ipea, 2007. V. 2.
Soares, S. O Ritmo De Queda Na Desigualdade No Brasil É Adequado? Evidências Do Contexto
Histórico E Internacional. Brasília: Ipea, 2008. (Texto Para Discussão, N. 1339).
Soares, S. A Desigualdade De Renda De 1995 A 2009 E Tendências Recentes. Rio De Janeiro: Centro De
Estudos Sobre Desigualdade E Desenvolvimento (Cede/Uff), 2011. (Texto Para Discussão, N. 51).
Soares, S.; Nascimento, P. A. M. N. Evolução Do Desempenho Cognitivo Do Brasil De 2000 A 2009
Face Aos Demais Países. Brasília: Ipea, 2011. (Texto Para Discussão, N. 1641).
Soares, S. Et Al. Conditional Cash Transfers In Brazil, Chile And Mexico: Impacts Upon Inequality.
Estudios Económicos, Número Extraordinário, P. 207-224, 2009.
Soares, S.; Sátyro, N. Análise Do Impacto Do Programa Bolsa Família E Do Benefício De Prestação
Continuada Na Redução Da Desigualdade Nos Estados Brasileiros - 2004 A 2006. Brasília: Ipea,
2009. (Texto Para Discussão, N. 1435).
Soares, S. Et Al. Os Impactos Do Benefício Do Programa Bolsa Família Sobre A Desigualdade E A
Pobreza. In: Castro, J. A.; Modesto, L. (Eds.). Bolsa Família 2003-2010: Avanços E Desafios.
Brasília: Ipea, 2010. V. 2p. 366.
Souza, P. H. G. F. Uma Metodologia Para Decompor Diferenças Entre Dados Administrativos E
Pesquisas Amostrais, Com Aplicação Para O Programa Bolsa Família E O Benefício De Prestação
Continuada Na Pnad. Brasília: Ipea, 2010. (Texto Para Discussão, N. 1517).
World Bank. Global Purchasing Power Parities And Real Expenditures - 2005 International
Comparison Program. Washington Dc: The World Bank, 2008.
.

NOTES

1. For more detailed analyses of the Brazilian economy in the second half of the 20th century, see, among several others,
Fishlow (1980), Sayad (1995), Abreu (2000) and Baer (2009).
2. Per capita GDP growth averaged 3.6 per cent between 2004 and 2008, also the highest since the late-1970s.
3. For more on the PNADs, see Martine et al. (1988), Médici (1988), Rocha (2003) and Martine (2005).
4. As in all other graphs, data points for 1994 and 2000 were obtained through linear interpolation.
5. For more details, see Ipea (2009).
6. See <http://portal.inep.gov.br/estatisticas-gastoseducacao-despesas_publicas-p.a._paridade.htm>.
7. Years of schooling vary from 0 (no schooling) to 15 (completed post-secondary education).
8. For a detailed historical analysis of the minimum wage in Brazil, see Dieese (2010).
9. States and municipalities can also set their own minimum wages as long as those wages are higher than the federal
one. Nevertheless, this has seldom been the case and the federal minimum wage remains the truly relevant one.
10. An example might make this point clearer: If the cumulative inflation was 5 per cent between years t-1 and t and real
GDP growth in year t-2 was 3 per cent, then the minimum wage adjustment in t should be 8 per cent.
11. In fact, these numbers understate the scope of the minimum wage benefits, as it is known that the BPC, alongside the
Programa Bolsa Família, is one of the social programmes that are heavily underestimated by the PNADs (Souza, 2010).
According to administrative data, the BPC had approximately 3 million beneficiaries in 2009.
12. For more details on the development of Brazilian Social Security and its institutional framework, see Ipea (2009).
13. For further information, see Ipea (2009) and Jaccoud, Hadjab & Chaibub (2010).
14. Earlier in 2011, President Dilma Rousseff announced that this cap would be increased to five children per family, but
this is yet to be implemented.
15. In 2011, all benefits were adjusted, although only the conditional ones had real gains.
16. As the PNADs do not explicitly identify the recipients of either programme, this paper used the well-tested ‘typical
values’ methods developed— among others— by Barros, Carvalho & Franco (2006) and Soares et al. (2010) to
disaggregate income transferred by the BPC and the Bolsa Família.
17. The over-bar is a simple average between the two data points.
18. Unfortunately, the PNADs do not discriminate between the Private and Public sectors’ Social Security. Nevertheless,
the two income components presented above— pensions tied to the minimum wage and pensions higher than the
minimum wage— work as a decent, albeit far less-than-perfect identification strategy: over 60 per cent of the private
sector pensions fall into the first group, whereas the public sector pensions are almost entirely concentrated in the latter.
19. In 2002, the PNADs featured slightly modified classifications for some variables. Even though the changes were
somewhat minor, we opted to use the 2002 data set— and not the 2001 one— in order to ensure perfect comparability.

International

Centre for Inclusive Growth

International Policy Centre for Inclusive Growth (IPC - IG)
Poverty Practice, Bureau for Development Policy, UNDP
Esplanada dos Ministérios, Bloco O, 7º andar
70052-900 Brasilia, DF - Brazil
Telephone: +55 61 2105 5000

E-mail: ipc@ipc-undp.org URL: www.ipc-undp.org

Poverty, Inequality and Social Policies in Brazil: 1995-2009

More Related Content

What's hot (20)

Viewers also liked (13)

Similar to Poverty, Inequality and Social Policies in Brazil: 1995-2009 (20)

More from UNDP Policy Centre (20)

Recently uploaded (20)

Poverty, Inequality and Social Policies in Brazil: 1995-2009