Inequality And Economic Growth In Brazil

2y ago
23 Views
2 Downloads
474.42 KB
41 Pages
Last View : 2m ago
Last Download : 2m ago
Upload by : Konnor Frawley
Transcription

Inequality and Economic Growth inBrazilBachelor's Thesissupervised by theDepartment of Economics at the University of ZurichProf. Dr. Fabrizio Zilibottito obtain the degree ofBachelor of Arts in EconomicsAuthor:Course of Studies:1Yves KellerEconomics

AbstractThis bachelor thesis gives an overview over Brazil's inequalityand provides the economic theory and latest ndings about the relationship between inequality and economic growth. In an empiricalpart the paper shows with newer data that a signi cant inequalitybased Kuznets' curve exists for the case of Brazil and that the levelof GDP alone can explain most of the variation in inequality. In asingle-country growth regression the paper nds evidence that education was the main contributor to economic growth in the last halfcentury and that the Gini coe cient is insigni cant when it comesto its in uence on growth in Brazil.2

Contents1 Introduction52 Inequality in Brazil82.1Inequality measures. . . . . . . . . . . . . . . . . . . . . . . . .82.2Facts and gures about Brazil's inequality . . . . . . . . . . . . .102.3Why is Brazil so unequal12. . . . . . . . . . . . . . . . . . . . . .3 Theoretical Channels on how growth can be a ected163.1Credit-market imperfections . . . . . . . . . . . . . . . . . . . . .173.2Redistribution programs . . . . . . . . . . . . . . . . . . . . . . .173.3Sociopolitical unrest. . . . . . . . . . . . . . . . . . . . . . . . .183.4Saving rates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .183.5Further channels3.6Net e ect3.7Situation in Brazil. . . . . . . . . . . . . . . . . . . . . . . . . . .18. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .19. . . . . . . . . . . . . . . . . . . . . . . . . .4 Empirical ndings on the e ect on growth19224.1Evidence of Kuznets curve . . . . . . . . . . . . . . . . . . . . . .234.2Cross-country regressions. . . . . . . . . . . . . . . . . . . . . .244.3Implications and problems . . . . . . . . . . . . . . . . . . . . . .255 Regressions265.1Data set . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .275.2Kuznets' curve in Brazil . . . . . . . . . . . . . . . . . . . . . . .285.2.1Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . .285.2.2Results. . . . . . . . . . . . . . . . . . . . . . . . . . . .295.35.45.5Determinants of Income Inequality . . . . . . . . . . . . . . . . .305.3.1Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . .305.3.2Results. . . . . . . . . . . . . . . . . . . . . . . . . . . .31E ect on Growth . . . . . . . . . . . . . . . . . . . . . . . . . . .325.4.1Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . .325.4.2Results. . . . . . . . . . . . . . . . . . . . . . . . . . . .32Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .346 Conclusion363

List of Figures1PPP GDP per capita in US Dollar for Brazil 1981-2009. . . . . .52Real GDP growth rate for Brazil 1999-2010. . . . . . . . . . . .63GDP per capita vs. Inequality. . . . . . . . . . . . . . . . . . .74Gini-coe cient on income inequality in Brazil 1981-2009.5Income distribution in Brazil 2009. . . .11. . . . . . . . . . . . . . . . .12. . . . . . . . . . . . . . . . . . . .11List of Tables1Income shares in Brazil 20092Education levels in Brazil 2010. . . . . . . . . . . . . . . . . . .133Data sources - Overview . . . . . . . . . . . . . . . . . . . . . . .284Data information . . . . . . . . . . . . . . . . . . . . . . . . . . .285Results from Kuznets curve regression. . . . . . . . . . . . . . .296Results: Determinants of Income Inequality . . . . . . . . . . . .317Results: E ect on Growth . . . . . . . . . . . . . . . . . . . . . .344

1 IntroductionBrazil's economic and political success in the last two decades, in which thecountry's GDP grew at an average rate of 3.2%, has fascinated researchers allover the world. Today, Brazil is the world's 8th largest economy (by PPP GDP)and has a GDP per capita of 11 600.1It is expected to have strong growthrates in the near future and is likely going to overtake even more countries interms of economic power. Brazil also experienced political success, getting tohost some of the world's biggest events in the future, e.g. the 2014 FIFA World' W ƉĞƌ ĐĂƉŝƚĂ ŝŶ h ŽůůĂƌ ;ůŽŐ ƐĐĂůĞͿCup and the 2016 Summer Olympic Games.WWW ' W ƉĞƌ ĐĂƉŝƚĂ ŝŶ h ŽůůĂƌϲϬϬϬzĞĂƌFigure 1: PPP GDP per capita in US Dollar for Brazil 1981-2009.Source: Penn World Table 2012Despite its recent economic success, Brazil has also one of the highest inequalities in the world. Its Gini index on income inequality of 51.9% is the 19thhighest in the world and the highest of all the 20 biggest economies in terms ofGDP. The data shows that inequality in Brazil has been relatively stable for along period of time and only began to decline substantially in the last decade.It is still a very high value of inequality which is supported by other inequalitystatistics as well: The income share held by the lowest 10% is a mere 1.1% while1 2011 est. (Source: CIA Factbook)5

ZĞĂů ' W ŐƌŽǁƚŚ ƌĂƚĞZĞĂů ' W ŐƌŽǁƚŚ ƌĂƚĞ ;ŝŶ йͿϴϳϲϱϰϯϮϭϬͲϭϭϵϵϵ ϮϬϬϬ ϮϬϬϭ ϮϬϬϮ ϮϬϬϯ ϮϬϬϰ ϮϬϬϱ ϮϬϬϲ ϮϬϬϳ ϮϬϬϴ ϮϬϬϵ ϮϬϭϬzĞĂƌFigure 2: Real GDP growth rate for Brazil 1999-2010Source: CIA Factbook2the highest 10% get 43.1%.The fact that Brazil has grown so fast and thus has become such a big economy in the last decades but still has one of the world's highest Gini coe cientsmakes it an interesting research topic to try to answer a question that has longbeen discussed by the best economic researchers: What e ect does inequalityhave on country's economic performance? Kuznets (1955) was one of the rstto present his theory to the audience. He argued that during a country's economic development, it rst goes through a period where inequality increasesand later declines.One possible reason might be the introduction of a newtechnology which rst favours only a few and therefore increases inequality butas more people adapt to the new technology inequality decreases again overtime.Williamson (1985), for example, showed that there is good evidence ofa Kuznets curve in England and Wales during 1823-1915. In a cross-sectionalapproach (Deininger and Squire, 1998), however, the evidence of a Kuznets'curve is less clear.The data shows that most countries with very high Ginicoe cients are in Latin America (Figure 3 in red dots), with Brazil as one ofthe most unequal.This thesis aims to improve the understanding of the relationship between2 2009 (Source: tradingeconomics.com)6

' W ƉĞƌ ĐĂƉŝƚĂ ǀƐ͘ ŶƚϲϬ ϬϬϬWWW ' W ƉĞƌ ĐĂƉŝƚĂ ;ƌĂƚŝŽ ƐĐĂůĞͿFigure 3: GDP per capita vs. InequalitySource: CIA Factbook (GDP data from 2011, Gini data latest possible)inequality and economic growth in a theoretical and empirical aspect and triesto give explanations for Brazil's enormous inequality and what e ect it has onits economic growth performance. Section 2 will rst present a few measureson how inequality can be assessed and what these numbers look like for Brazil.It also compares Brazil in a cross-country comparison and tries to explain whyBrazil is such an unequally distributed country. Section 3 will explain possibletheoretical channel on how inequality can actually a ect growth.Followingthese theoretical aspects, the paper tries to quantify the importance of thesepossible channels for the case of Brazil.Section 4 will give an overview overthe latest ndings on the evidence of a Kuznets curve and on cross-countrygrowth regressions that include inequality variables.In an empirical Section5 the bachelor thesis proves with better and newer data the evidence of aninequality-based Kuznets curve in Brazil for the last 50 years.Results willsupport the existing research which has Brazil as one of only a few countrieswith such a relationship. Further, the section will carry out regressions aboutother possible determinants of inequality apart from the country's GDP and willpresent growth regressions for the single country of Brazil. Dummy variablesfor democratic circumstances as well as for the latest political party in charge,i.e. the Worker's Party, will be included in the regression. Section 6 concludes.7

2 Inequality in BrazilAs described in the Introduction, Brazil is infamous for its extreme inequalitywith a Gini index on income inequality of 51.9%. However, there are more thanone method to describe a country's inequality and some of them will be explainedin the rst subchapter. In a second part the results for Brazil are presented andput in a world wide context. In a nal subchapter, I will summarize the ndingsin economic literature on why Brazil is so unequal.2.1 Inequality measuresFirst and foremost it has to be said that there is no such thing as theright inequality measurement. The question which method should be chosen isdependable on what speci c point the author wants to focus or on how suitablethe method is in explaining the given problem. In addition, most of the timeresearchers have to rely on the available data collected by national and regionalgovernments. Calculating inequality is something not only done by economistsbut also by scientists in many other elds, ranging from social sciences to naturalsciences.PropertiesDespite di erent methods in calculating inequality, there are four properties that scientists can agree on (Cowell and Victoria-Feser, 1996): (1) it shouldmake no di erence in the calculation which person owns a speci c income share,(2) richer countries should not by construction be labelled more unequal (scaleindependence), (3) a higher population should have no e ect on the measurement (population independence) and (4) a transfer from a richer person to apoorer person while still preserving the order of income ranks should decreaseand not increase inequality.Gini indexBy far the most famous and most used method in measuring inequality isthe Gini index. De ning the area between the the line of perfect equality andthe Lorenz curve as A and the area under the Lorenz curve as B (see also Figure5 later in this chapter for the case of Brazil), the Gini index is given by:Gini Gini A0.5A(A B) 2A 1 2B8

If the Lorenz curve is given by a functiony f (x),the Gini index can befound like this:Gini 1 2 10f (x)dxIn reality, though, the Lorenz curve is almost always unknown and oftenonly certain points on the curve, i.e. income shares of certain population shares,are aquired in surveys. This makes the calculation of the area B easier but itmust be said that those results represent only approximations. Barro (2000),for example, uses that kind of approximation, a Gini index based on incomequintile-shares rst introduced by Theil (1967). A Gini index can take valuesfrom 0 to 1 with 0 being a total equal and 1 a total unequal society. For furtherresearch and more advanced methods one can have a look at Theil (1969),Ogwang (2000), Giles (2004) or Deaton (1997) who established a very simpleformula for calculating the Gini index with only the mean of the distribution,the population and the income shares.Hoover indexThe Hoover index, also called Robin Hood index, is an other measure ofinequality that is used by scientists (Hoover, 1936). It describes how much ofa country's or region's income had to transferred from the rich segment to thepoor one in order to have a community that is totally equal. Graphically, it isrepresented by the largest di erence between the total equality curve and theLorenz curve. As with the Gini index, the Hoover index can take values from 0to 1 with 0 being a total equal community.Further indexesThe Theil index (Theil, 1967) and the Atkinson index (Atkinson, 1970) aretwo more complex indexes which are both based on entropy indexes origing ininformation theory. These indexes are preferred when subgroup inequality, i.e.inequality among income share groups, should also be considered.However,for most economic research and also for this paper, such indexes are not oftenmade use of and the Gini index is commonly preferred due to its simplicity and3availability in statistical research.3 For further study: Atkinson and Bourguignon (2000) and Sen and Foster (1996)9

Income-share and ratiosAn other well used method in describing a community's inequality is to takea look at income shares.Those shares explain how much a certain share ofpopulation gets from the total income. Usually, one is most interested in thehighest- and lowest-income shares, usually with values that range from 0.1%to 20% of the population. For comparing those shares' income di erences, researchers often build ratios to explain by how many times one group is bettero than an other. Therefore, it's usually higher income groups over lower income groups. Those income shares can also be compared to mean or medianvalues to get more comparable results. One huge disadvantage compared to theabove stated indexes might be that ratios do not provide absolute measures ofinequality.After summarizing the most important ways how to calculate inequality, oneimportant point should not be forgotten and starts before choosing an adequatemethod in analysing the inequality.That is the answer of which inequalityshould be measured. An economic researcher has to distinguish between incomeand wealth inequality; and if he chooses income inequality, does it include capitalgains or not. He also has to decide whether to look at individuals or householdand if he wants to focus on inequality before or after tax. The di erent methodsfor putting a number on inequality explain practically the same thing, but cansometimes vary a bit in their levels.For high inequalities, the Theil index islarger than the Hoover index, while for low inequalities, it is the other way round(Deaton, 1997). In general, most research done on the topic of inequality usesa second inequality measure to test their results' robustness. Barro (2000) usedthe Gini coe cient in his main regressions and tested his ndings against thelowest and the highest income-quintile share to con rm his original regressionresults.2.2 Facts and gures about Brazil's inequalityIn 2012, Brazil was ranked 19th in world in income inequality measured bythe most popular inequality index, the Gini coe cient. Its value was measuredat 51.9%, approx. two points down since 2009. Over the course of the last halfcentury, Brazil's Gini coe cient has been fairly constant with a peak of 63% inthe late 80's. There has been a sharp decline in the beginning of the 90's, butthe Gini coe cient stayed close to 60% until the new century. Since 2002, thedecline of income inequality has accelerated, but is today still at closely over 50%10

ŝĐŝĞŶƚ ŝŶ e 4: Gini-coe cient on income inequality in Brazil 1981-2009.Source: Index Mundi 2012Income group Income sharelowest 20%3.2second 20%7.1third 20%11.9fourth 20%19.4highest 20%58.5Table 1: Income shares in Brazil 2009Source: tradingeconomics.com 2012(Figure 4). This makes Brazil, as seen in Figure 3 in the Introduction, the mostunequal country of the 20 biggest economies in the world, with a signi cantlyhigher Gini coe cient than the current 6 larger economies, measured by PPPGDP (USA 45%, China 48%, India 36.8%, Japan 37.6%, Germany 27%, Russia442%) .Graphically, Figure 5 shows Brazil's income distribution as of 2009 whileTable 1 presents the values for each income group. It shows that in the lowestincome share, 20% percent of the population have to live with only 3.2% of the4 2011 (Source: CIA Factbook)11

ƵŵƵůĂƚŝǀĞ WĞƌĐĞŶƚĂŐĞ ŽĨ /ŶĐŽŵĞ/ŶĐŽŵĞ ĚŝƐƚƌŝďƵƚŝŽŶ ŝŶ ƌĂnjŝůϭϬϬϴϬϲϬ ŽƌĞŶnj ĐƵƌǀĞϰϬdŽƚĂů ĞƋƵĂůŝƚLJϮϬϬϬϮϬϰϬϲϬϴϬϭϬϬ ƵŵƵůĂƚŝǀĞ WĞƌĐĞŶƚĂŐĞ ŽĨ WŽƉƵůĂƚŝŽŶFigure 5: Income distribution in Brazil 2009Source: tradingeconomics.com 2012total income generated in the country. In contrast, in the richest group, 20% ofthe population accounts for 58.5% of all the income. Historic data from the lastdecade shows that the lowest income group has increased its income share from2.7% in 2000 to 3.2% in 2009, a rise of 15.6%. On the other hand, the highestincome group has lost some of its share, dropping 6.6% from 62.3% to 58.5% inthose nine years.2.3 Why is Brazil so unequalLooking at the bare numbers in the preceding subchapter the question ariseswhat could possibly be the reasons for Brazil's enormous inequality. Income inequality exists because people di er in all kind of aspects that are relevant tothe income they achieve. What usually contribute most to income inequalityare di erences in human capital, i.e. education and health. Those di erences inhuman capital will result in di erent returns on the labour market. The moreunequal the distribution of human capital and the higher the return to education, i.e. a skill premium, the more unequal will also the income be distributed.Skill premiums are driven by supply and demand on the labour market andon productivity a worker can achieve (Weil, 2012). However, income inequality can also have di erent sources: there can be geographical di erences, e.g.12

Level of schooling % attained (% completed)No schooling13Primary schooling39 (26.2 completed)Secondary schooling40.3 (25.1 completed)Tertiary schooling7.3 (5.2 completed)Table 2: Education levels in Brazil 2010Source: Barro and Lee (2010)between urban and agricultural regions, gender- or race-discrimination, technology changes or international trade increases that favour the already skilled andhigh-earning share of the population or the role the government plays when itcomes to inequality, mainly through its redistribution policies it will implement.Di erences in education and in the return to itThe data on education from Barro and Lee (2010) shows a very unbalanceddistribution of human capital in Brazil.The numbers in Table 2 show thepercentages of the population aged 25 or older who have attained (completed)primary, secondary and tertiary level of education in the year 2010.Bourguignon et al.(2002) analysed the impact of such an unequal distri-bution of education, comparing it to other countries (Mexico, United Statesand Colombia).They found that di erences in returns to schooling explain,together with work-experience, approximately 40% of the di erence in inequality between Brazil and the United States. This proves to be one of the mostimportant reasons why Brazil faces these kinds of inequalities on income.As seen in Table 2, Brazil has an extremely skewed allocation of humancapital.This will translate via the labour market into even higher incomeinequality in Brazil as opposed to somewhere else, because skill-premiums inBrazil are higher compared to the countries also analysed in Bourguignon et al.(2002). According to the paper, steeper returns to education constitute 20% of5the excess inequalitycompared to the United States.Geography & DiscriminationFurther, there can be income di erences amongst di erent regions in a country. According to the Instituto Brasileiro de Geogra a e Estatística, hereafter5 additional inequality when compared to an other country13

IBGE, which categorizes the country into 5 big regions, there are minimal differentials in Gini coe cients among those regions. The centre-west region, including the capital Brasília, features the highest income inequality with a Giniindex of 55.4%.The lowest Gini value is found in the south region (48.9%),which is also the smallest of the 5 big regions in Brazil.Geographical di erences in Gini coe cients can also be looked at from another perspective: One can distinguish between urban and rural areas. In 2005,the urban Gini of Brazil was 60%, while the rural Gini was 54%, both beingsigni cantly over the average of other Latin American countries (50% for both6urban and rural Gini).Inequality can also have its sources from intended or unintended discrimination, either in the labour market or directly through the government, e.g.favouring a speci c group of the population. Two forms o

inequality and economic growth in a theoretical and empirical aspect and tries to give explanations for Brazil's enormous inequality and what e ect it has on its economic growth performance. Section 2 will rst present a few measures on how inequality

Related Documents:

growth will be discussed. Concepts on both regional income inequality and local economic growth will follow. 2.1 Theories on Income Inequality: Does Harm Growth Stiglitz (2012) argues inequality slows economic growth. According to Stiglitz, inequality weakens aggregate demand f

Measuring economic inequality Summary Economics 448: Lecture 12 Measures of Inequality October 11, 2012 Lecture 12. Outline Introduction What is economic inequality? Measuring economic inequality . Inequality is the fundamental disparity that permits one individual certain

impact of economic growth, with a range of socio-economic indicators, the Survey highlights that both economic growth and inequality have similar relationships with socio-economic indicators. Thus, unlike in advanced economies, in India economic growth and inequality converge in terms of the

economic growth on income inequality. The first argument is that inadequate redistributive policies and the increase in inequality that accompany economic growth lessen the potential benefits of economic growth to the

inequality transmits to growth, it enables inequality to have growth or level effects depending 1 Recent studies that investigate the relationship between inequality and growth include Barro (2000), Forbes (2000), Panizza (200

A Model of Gender Inequality and Economic Growth This paper introduces a model of gender inequality and economic growth. The model is calibrated using microlevel data of Asian economies, and numerous policy experiments are conducted to investigate how various aspects of gender inequality are related t

economic growth, we also use human capital inequality measures. In fact, the role played by human capital inequality on economic growth is present in most of the models that analyze the effect of inequality on

awards will be separately funded from outside the 0.3 allocation per eligible Trust consultant using the same criteria as set out in 2.9. 4. Eligible Consultants under Investigation 4.1 If a consultant who is the subject of a formal investigation, including a professional advisory panel, chooses to submit an application for CEAs, his/her application will be considered in the usual way by the .