BrazilIncomeDistribution and Economic Growth

INCOME DISTRIBUTION AND
ECONOMIC GROWTH: THE CASE
OF BRAZIL


Ms. Camila Matins-Bekat
University of Denver, USA
Kishore G. Kulkarni
Metropolitan State College of Denver, USA


ABSTRACT

In a perfect world, it will be great if the growth can be distributed equally and all poverty is
removed simultaneously. Unfortunately, as Simon Kuznets has pointed out, economic growth is
rarely distributed equally, in fact, according to him, economic growth initially leads to higher
income inequality of income, as some sectors grow faster and some do not grow at all. In this
paper, we apply the Kuznets hypothesis to the Brazilian case. Our test shows that in case of Brazil,
the Gini coefficient has in fact behaved as Kuznets predicted.

JEL Classifications: O54, O40
Keywords: Income distribution and economics growth, country studies: Brazil
Corresponding Author’s Email Address: [email protected]

INTRODUCTION
As inequality applied to income and economic inequality, is a very complex issue.
Economists, sociologists, and political scientists are only a few of the researchers
concerned with economic inequality. Is all inequality equal in its effects? Alternatively,
do certain types of inequality spur growth while others damage it? Concrete answers to
questions regarding the relationship between inequality and economic growth have not
been found (see Introduction: Seligson and Passé-Smith, 2003). More specifically,
development economists have been trying to understand the reasons why economic
inequality is higher in countries that grow fast. Is economic growth the culprit?
These and other related questions have been the focus to study for several
prominent economists. Simon Kuznets developed the most famous hypothesis of the
relationship between inequality and economic growth. He argued that as a country
develops, income inequality initially increases, and only after some time, it declines. In
the first section of this paper, we will briefly discuss the social, political, and economic
impact of income inequality. The second section is a review of the literature for and
against Kuznets‟ hypothesis. In section three, we will apply Kuznets‟ hypothesis to the


342
case of Brazil using time-series data. Suggestions for ameliorating income inequality in
Brazil are discussed in the conclusion.

IMPACTS OF INCOME INEQUALITY

Social
Gottschalk and Justino (2006) highlight several important studies regarding the social
impact of income inequality and present the argument that high inequality may
deteriorate stocks of human capital when associated with high illiteracy and poor health.
Ribero and Nunez showed that there is a negative correlation between disability and
stature and the ability to earn an income. In other words, an individual‟s ability to earn
money decreases as their nutritional status deteriorates (Gottschalk and Justino, 2006).
The evidence suggests that another social impact of income inequality is the
positive relationship between high inequality and forms social and political conflict
(Lichbach, 1989 in Gottschalk and Justino, 2006). If the poorest 20% of a society control
less than one percent of the national income, their hardships are exasperated. Lack of
access to education, healthcare, social security and jobs only fuel the sentiment that the
government is not doing its share to alleviate their suffering. Once the tolerance threshold
for inequality is reached, society is faced with what Hirschman and Rothschild (1973)
called the Tunnel Effect (Ray 1998). The implications of Hirschman‟s tunnel effect
hypothesis on economic growth policy are significant. If the tunnel effect is present in a
weak society, then an economic policy focused on aggregate growth would not be a
prudent choice. Instead, the society should choose a policy that would address growth
and distribution simultaneously (1998).

Political
Illiterate and poor, the individuals at the lowest ranks of society are often excluded from
the political arena either because they cannot afford to vote or due to voter corruption.
The latter takes the form of candidates offering small amounts of money to the poor in
return for their „support.‟ Thus, public policies in high inequality countries regularly
exclude the poor (Gottschalk and Justino, 2006). In general, Latin American public
policies tend to be the most regressive in the world, which means that the policies in
pursuit of fiscal goals do not prioritize the needs of the poor (Weyland, 1996 in
Gottschalk and Justino, 2006). The most influential individuals in the political arena are
therefore the rich minority. The weight carried by traditional families and businesses
transfer to government policies. A result of this is that developing countries with high
levels of inequalities often have a low tax rate in proportion to the GDP (2006).

Economic
Broadly speaking the economic impact of inequality is widespread and considerable. It
affects the demand structure of the lower and middle class. If the lowest class is
perpetually at a state of subsistence, lacks access to education, healthcare, social security
benefits and has no political influence they will remain in poverty unable to consume a
higher percentage of goods and services. According to Murphy et al. [(1989) in
Gottschalk and Justino, 2006)] “decreases in income inequality imply a wealthier middle


343
class, (enlarged by those coming out of the poorer classes), which are the most significant
consumers of manufactured goods.”
The saving rate is significantly affected at low incomes. According to the
Harrod-Domer Model of Economic Growth, the availability of saving is the key to
economic growth. A lower the saving rate results in less money available for investment
therefore yielding a lower rate of growth (see Ray, 1998). The creation of the larger
middleclass will also increase the saving rate. In general, improved opportunities for the
poor in the form of reduced income inequality will lead to a higher volume of
consumption thereby expanding the internal market, a higher savings rate, and ultimately
a better position for economic growth.

LITERATURE REVIEW

Simon Kuznets
Simon Kuznets‟ Inverted-U hypothesis was based on relatively scant cross-sectional data
from 1948-1950 from the United Kingdom, and the United States for the developed
countries and Puerto Rico, India and Ceylon for the less developed countries. The data
indicated that income distribution in these developed countries was somewhat more equal
than in these less developed countries (Kuznets, 1955). His findings are represented
below.
Table 1. INCOME INEQUALITIES OF THE COUNTRIES

Country Lower 3 quintiles (%) Top quintile (%)
India 28 55
Ceylon 30 50
Puerto Rico 24 56
United States 34 44
United Kingdom 36 45
* Data from Kuznets, 1955.

Kuznets points out that “Much is to be said for the notion that once the early
turbulent phases of industrialization and urbanization had passed, a variety of forces
converged to bolster the economic position of the lower-income groups within the urban
population” (1955; 17). The forces Kuznets is referring to include a better adaptation and
organization of the recent immigrants and an increase in their ability to deal with the
hardships of urban life. With time, Kuznets assumes that the lower income class will be
able to exert enough political influence to bring about changes in legislation in favor of a
more equitable distribution. However, this is not the case in the developing world (see
Weyland, 1996). Kuznets also points to the “dislocating effects of the agricultural and
industrial revolutions,” coupled with rapid population growth in the cities due to a rise in
birth rates and a decline in death rates, and concludes that the early phases of economic
growth bring about a more unequal size distribution followed by a period of stabilization
ending with in a more equal size distribution during later stages of growth after the “new
industrial system had shattering effects on long-established pre-industrial economic and
social institutions” (1955; 18). This conclusion is what we now call the Kuznets‟
Inverted-U Hypothesis.


344
FIGURE 1. INCOME INEQUALITY OVER TIME


Oshima states that “the major determinant of the dispersion of quintile shares between
countries is the weight of the farm or rural sector in the total economy” and that other
factors have a minimal impact (1962; 442). However, he qualifies this finding by
pointing out that the influence of heterogeneous factors such as history, geography,
politics, religion and race on income inequality render the conclusion that developing
countries have a higher income inequality than developed countries. Kravis (1960)
confirms “Distribution of income tends to be more equal the longer and the more
thoroughly the country has been exposed to the processes of economic and social change
associated with the idea of industrialization” (1960; 409). The greater initial inequality is
the result of a partial integration of the population in new sectors of the economy. As the
population becomes more fully integrated, the economic differentiation between sectors
declines.
In the case of Brazil, with its high number of Afro-Brazilian population, studies
by Pastore (1982 and 2000) confirm the non-white individuals have a significant
disadvantage when it comes to occupational mobility. With only 47.8% of blacks
elevating themselves to a higher social status than their parents compared to 55.8% of
whites (2000; 89). The implication being that non-whites often remain in the lowest
income brackets while the income of whites increases. From the data in his study, Kravis
concludes, “racial heterogeneity tends to increase the degree of inequality” (1960; 413).
A larger study by Paukert (1973) of 56 countries, 40 of which were developing
countries showed that the degree of inequality is associated with level of GDP per capita.
The sharpest increase in inequality resulted from a move from the lowest income level
(GDP per capita under $100) to the second lowest level with a GDP per capita between
$101- $200. Although he observed that inequality continued to increase until the $2000
GDP per capital level, the highest increases occurred from the first to shifts in income
level. His study also showed that after countries reach the $2000 GDP per capital level
inequality decreases, confirming Kuznets Inverted U Hypothesis. Ahluwalia (1976)
using a different data set also came to the same conclusion.

Problems with Kuznets’ Inverted U Hypothesis
A common complaint surrounding Kuznets‟ work revolves around the availability of
data. As Ray (1998) highlights, inequality data from countries that have already
completed their inverted-U path is not available and that in conducting a cross-section
Time


345
study it is important to control for inter-country variation. This fact is also supported by
Fields (1989) who points out that while Kuznets makes predictions that span several
generations, available data was for a period of ten years on average. “There might be
evidence of trends, but the variation within those trends may obscure longer-term
movements” (169). Pointing our another complication, Fields (1989) highlights the fact
that the Gini does not incorporate quality of life nor can it account income generated in
the informal sector of a society, which tends to be much larger in developing countries.
He tested several hypotheses regarding the relationship between economic growth and
inequality. Using time-series data and looking at both the Gini and Lorenz curves,
Fields found that inequality increased in just about the same number of countries as it
decreased (1989;171).
Fields also claims that “growth tends to raise inequality in low-income
countries and to reduce inequality in high-income countries” (1989; 176). He tested this
hypothesis by using inter-temporal data, which he claims is the correct type of data to
use. The study reveals that inequality increased nearly as often as it decreased, 48% and
42% respectively (176). Fields concluded that there is no apparent trend for countries to
follow the Kuznets Inverted U hypothesis. We now investigate the application of Kuznets
Hypothesis in the context of Brazilian economic growth with a special focus on the
hyperinflation period of 1989 to 1993 and the stabilization efforts of the 1994 Real Plan.

APPLICATION OF KUZNETS’ INVERTED U TO BRAZIL

It is of no surprise the income distribution in Brazil is among one of the most unequal in
the world. Using the Gini Coefficient as the measure of income inequality, in 1989
Brazil ranked second highest in the world (.625) only falling behind Sierra Leone
(Ferreira, Leite & Litchfield, 2006). It currently ranks tenth in the world. Only two Latin
American countries have a Gini higher than Brazil‟s: Guatemala and Bolivia (2006).
Brazil‟s level of inequality is more in line with that of Sub-Saharan African countries
such as Botswana, Lesotho, South Africa, and Zimbabwe. Pastore (1982 and 2000) has
tracked social mobility in Brazil. His findings show that while Brazil has a higher volume
of social mobility than Spain, France, the U.S. and Poland, most of the mobility, 62.3%,
is accounted by a move from the “inferior-low” (manual rural laborer) status to the
“superior-low” (unspecialized manual urban laborer) status and from “superior-low” to
“inferior-medium” (semi-qualified urban laborer –such as housekeeper and street vendor)
status. The Word Bank Country Study (2004) reveals that in Brazil skin color accounts
for 12% of income inequality. This fact can be attributed to the differences in education
level, access to healthcare and wages between whites and non-whites. The Study also
reports that regressive public transfers such as retirement pensions for government
employees, inequitable distribution of education and high wage differentials by skill level
account for 39%, 29% and 32% of Brazil‟s excess inequality. However, since 1994,
social programs such as Bolsa Escola, have contributed to income equalization.

The Numbers
Now we look at the data available to determine whether the relation between Brazil‟s
economic growth from 1980 to 2005 and its inequality follows Kuznets‟ Inverted U
Hypothesis. While the focus of the data is from 1980, it is important to note that during


346
the Brazilian Economic Miracle (late 1960s and early 1970s) when GDP percent growth
was at nearly 10% per annum, there is little controversy as to whether Brazil‟s inequality
increased. The Gini coefficient rose from .50 in 1960 to .57 in 1970 and continued to
increase until 1976 before declining from 1977 to 1981 (Leite, Ferreira, & Litchfield,
2006;2). A study conducted by Fishlow (1972) confirms, “The conclusion that inequality
has increased over the course of the decade accordingly seems correct, if lamentable”
(399).
FIGURE 2. ANNUAL GDP GROWTH
1961-2006 (PERCENTAGE)


Source: World Bank Quick Query available at http://ddp-ext.worldbank.org

TABLE 2. ECONOMIC GROWTH AND GINI COEFFICIENT 1980-2006

Year GNI per capita (current US$) Gini Coefficient
1980 2190 .59
1982 2000 .58
1984 1620 .58
1986 1790 .58
1988 2250 .60
1990 2700 .60
1992 2780 .58
1994 3050 .62
1996 4470 .59
1998 4880 .59
2001 3290 .59
2002 3050 .58
2004 3220 .57
2005 3880 .56
Table adapted from Social-Economic Database for Latin America and the Caribbean (SEDLAC) available at
http://www.wider.unu.edu/research/Database/en_GB/income-distribution-links/, Instituto Brasileiro de Geographia e Estatistica (IBGE) available at
www.ibge.gov.br and World Bank Quick Query available at http://ddp-ext.worldbank.org


347
FIGURE 3. BRAZIL: GINI COEFFICIENT 1959-2005


Source: Franko, P., (2003) The Puzzle of Latin American Economic Development and Social-Economic Database for Latin America and the
Caribbean (SEDLAC) available at http://www.wider.unu.edu/research/Database/en_GB/income-distribution-links/. Graph generated by author.


GRAPH 3. BRAZIL: GINI COEFFICIENT 1980-2005 – A CLOSER
LOOK

Source: Ibid


In analyzing the Gini Coefficient over time, we can safely say that in the case of Brazil
inequality did increase during the initial stage of development, and that by the end of


348
1994 to 2005 it did indeed decline. From 1994, when the Gini Coefficient was at its peak
(.62) to 2005 when it was at its lowest (.56) there was nearly a 10% decline. From graph
3, we clearly observe that income inequality did decrease from 1989 to 2005.

I ncrease in I nequality: The debt and inflation crises
We see from the above table that GNI per capita during the early 1980s followed a
downward trend and picks back up toward the end of the decade. This initial decline can
be attributed to what economist call the “lost decade” when many of Latin American
countries were experiencing a debt and inflation crises. The debt crisis was a result of
careless borrowing and irresponsible spending habits during military rule (1964-1985).
An increase in interest rates in 1979 by the U.S. Federal Reserve Bank also
increased the interest payments on past loans (most of Brazil‟s debt was denominated in
dollars) (Franko, 1999). In order to cover its debts Brazil simply printed more money
resulting in hyperinflation. During the latter part of the 1980s, inflation rates were over
1000% (Font, 2003). As the economy scrambled to maintain afloat, funding for social
programs suffered. This, along with the inherent deterioration of incomes due to the debt
and inflation crises caused the income inequality to increase during the last three years of
the decade (Ferreira, Leite & Litchfield, 2006).

Decrease in I nequality: Stabilization
The decline in inequality that followed from 1994 to 2005 can be attributed to a series of
policies and reforms enacted by President Cardoso. Three previously failed stabilization
plans created a great deal of pressure and speculation as to whether Cardoso‟s plan would
be successful. The new currency, the Real, introduced in 1994, was pegged one-to-one to
the dollar. Cardoso saw high government spending and large fiscal imbalances as the
root causes of hyperinflation. To this end, he pursued structural and fiscal adjustments.
By 1995 fiscal balance was possible because of budget cuts and new measures taken
against tax evasion which is estimated to have cost the Brazilian government anywhere
from $25-30 billion a year. Second, overcoming of economic stagnation was
accomplished by a trade surplus from exports from 1993-95, increase in hard currency
reserves and of course, the decrease in inflation. By 2002, the strong currency, which
spurred growth in the import sector, turned the trade surplus into a deficit and the
decision to devalue the currency could no longer wait. The success of the Plano Real
hinged on keeping inflation down and for this Cardoso focused on having sound
monetary and fiscal policies. In all, the Plano Real worked.

TABLE 3. INFLATION, 1993-2001, IN PERCENT

1993 1994
(J-J)
1994
(J-
D)
1995 1996 1997 1998 1999 2000 20001 20002
a

2,564 258.7 16.8 18.8 9.4 6.1 1.0 14.3 7.4 9.3 10.7
Source: Font, M., (2003) Transforming Brazil; a. Annual inflation until September

From 2001 to 2005, Brazil‟s Gini saw a decline of 4.6% from .593 to .566, one
of the fastest rates in world (Paes de Barros et al., 2007). During this time frame the


349
income proportion of the poorest 20% increased by one-half of a percent. Brazil still has
a long way to go.

TABLE 4. PER CAPITA INCOME BREAKDOWN BY PERCENTILE AND GINI
COEFFICIENT BETWEEN 2001 AND 2005

Deciles 2001 2002 2003 2004 2005
Variation
2001-2005
First .69 .79 .78 .87 .91 .22
Second 2.36 2.55 2.56 2.79 2.87 .51
Third 4.85 5.12 5.17 5.57 5.68 .83
Forth 8.24 8.55 8.72 9.25 9.40 1.17
Fifth 12.7 13.1 13.3 14.0 14.2 1.52
Sixth 18.5 18.9 19.4 20.2 20.5 1.95
Seventh 26.1 26.5 27.1 28.1 28.4 2.28
Eighth 36.6 37.0 37.7 38.8 39.0 2.44
Ninth 52.8 53.2 54.0 55.0 55.0 2.17
Gini Coefficient .593 .587 .581 .569 .566 -4.61%
Source: Table adapted from Paes de Barros et al., 2007

Ferreira, (200) suggest the following explanation: “the decline in inequality
between educational sub-groups which appears to be driven by a persistent reduction in
the average returns to schooling in Brazil…income differences between the country‟s
urban and rural areas have fallen dramatically…..[and the] potential decline in racial
inequalities” (Ferreira, Leite & Litchfield, 2006;31).

CONCLUSION

Despite debate over whether Kuznets‟ hypothesis was founded proper use of available
cross-sectional data, the application of the inverted U to the case of Brazil using time-
series data reveals that inequality in Brazil does indeed fit an inverted U path. Hence in
early days the Brazilian economy experienced the increased inequality of income but in
the long term this inequality has subsided. Whether this accelerated downward trend in
inequality will persist, remains to be seen. However we can conclude that some
governmental policies have been successful. Expansion of income substitution programs
such as Bolsa Escola and Bolsa Familia to create an incentive for poor families to send
their kids to school rather than sending them to work menial jobs have seen some positive
results. Lastly, continued pension reform for government sector employees can benefit
the reduction in income inequality. In general, however, our main observation is that the
economic growth by itself is rarely equitable. To make it equitable, if at all that is an
objective of the government, special policy actions have to be adopted.







350
REFERENCES

Ahluwalia, M.S., (1976). “Income Distribution: some stylized facts,” American
Economic Review 66,128-35.
Paes de Barros, R., et al., (2007). A Queda Recente da Desigualdade de Renda
no Brasil.
Instituto de Pesquisa Economica Aplicada (IPEA). Available at
www.ipea.gov.br
Ferreira, Francisco H.G., Leite, Phillippe G. and Litchfield, Julie A., "The Rise
and Fall of Brazilian Inequality, 1981-2004" (March 1, 2006).
World Bank Policy Research Working Paper No. 3867 Available at SSRN:
http://ssrn.com/abstract=922993
Fishlow, Albert “Brazilian Size Distribution of Income.” American Economic
Review (62).
Fields, G.S., (1989). “Changes in Poverty and Inequality in Developing
Countries,” The World Bank Research Observer, 4(2), 167-185.
Fields, G.S., (1977). “Who benefits from Economic Development: A
reexamination of the Brazilian Experience,” American Economic Review 1977
Font, M.A., (2003). Transforming Brazil. Rowman & Littlefield Publishers:
Lanham, Maryland.
Franko, P. (1999). The Puzzle of Latin American Economic Development.
Rowman & Littlefield Publishers: Lanham, Maryland.
Kravis, I. (1960). “International Differences in the Distribution of Income,” The
Review of Economics and Statistics, 42(4), 408-416.
Kulkarni, K., (2006). Readings in International Economics: Selected writings of
Prof. Kishore G. Kulkarni. Second Edition, Serials Publications: New Delhi.
Kuznets, S. (1955). “Economic Growth and Income Inequality,” American
Economic Review 45, 1-28.
Gottschalk, R. & Justino, P. (eds.) (2006). Overcoming Inequality in Latin
America: Issues and challenges for the twenty-first century. Routledge: New York.
Oshima, H.T. (1962). “The International Comparison of Size Distribution of
Family Incomes with a Special Reference to Asia,” Review of Economics and Statistics
54, 439-445.
Papanek, G.F., and O. Kyn (1986). “The effect on Income Distribution of
Development, the Growth Rate and Economic Strategy,” Journal of Development
Economics 23, 55-65.
Pastore, J. (2000). Mobilidade Social no Brasil. Makron Books do Brasil
Editora Ltda: Sao Paulo, Brazil.
------------ (1982). Inequality and Social Mobility in Brazil. University of Wisconsin
Press: Madison, WI.
Paukert, F. (1973). “Income Distribution at Different Levels of Development: A
Survey of Evidence,” International Labour Review 108, 97-125.
Ray, Debraj (1998). Development Economics. Princeton University Press:
Princeton, NJ.
Sen, A. K. (1973). On Economic Inequality. Oxford: Clarendon Press.


351
Seligson, M. & Passé-Smith, J., (2003). Development and Underdevelopment:
the political economy of global inequality. Lynne Rienner Publishers Inc.: Boulder, CO.
Shiller, Robert. (2007). Is Inequality Destabilizing? Economist‟s View.
Retrieved on July 15, 2008 from http://economistsview.typepad.com/economistsview
/2007/02/in_inequality_d.html
World Bank, (2004). World Bank Country Study: Inequality and Economic
Development in Brazil. The World Bank: Washington D.
Copyright of Journal of Developing Areas is the property of Tennessee State University, College of Business
and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright
holder's express written permission. However, users may print, download, or email articles for individual use.