Mapping poverty, inequality and the new middle class in Brazil

Texto

(1)Mapping Poverty, Inequality and the New Middle Class Progress in Brazil (With special reference to Rio 2016 & 2014 World Cup cities). Coordination: Marcelo Cortes Neri 7th October 2009 - Version 1.0. 1.

(2) The published articles are entirely the sole responsibility of its author. His opinions here expressed do not reflect necessarily the viewpoint of the Getulio Vargas Foundation.. Mapping Poverty, Inequality and the New Middle Class Progress in Brazil (with special reference to Rio 2016 the 2014 World Cup cities) / Marcelo Côrtes Neri (Coord.). - Rio de Janeiro: FGV/IBRE, CPS, 2009. [130] p.. 1. Brazil 2. Inequality 3. Poverty 4. Social Mobility 5. New Middle Class 6. Rio de Janeiro 7. Olympic Games I. Neri, M.C. © Marcelo Neri 2009. 2.

(3) Centro de Políticas Sociais Instituto Brasileiro de Economia Fundação Getulio Vargas. Coordination: Marcelo Côrtes Neri Luisa Carvalhaes Coutinho de Melo. CPS staff: Samanta dos Reis Sacramento André Luiz Neri Ana Beatriz Urbano Andari Lucas Moreira Ana Lucia Salomão Calçada. Art by: Marlus Pires. 3.

(4) Table of Contents 1. Preface 2. Introduction 3. Rio 2016: a shock of progress? 4. General Brazilian background 5. Income distribution 6. Reasons for change: inequality; mean income 7. Brazilian income 8. Social draw 9. Mapping Brazilian Social Progress (income-based measures) 10. Objective 11. The Geography of Poverty 12. Real poverty decrease 13. Local contribution to poverty decrease 14. Rise of class ABC 15. Emergence of class ABC 16. Inequality reduction 17. Source of income 18. Capitals of income 19. Class ABC – what happened since September 15th 2008? 20. Appendix I – Maps of Social Evolution 21. Appendix II – The Principal component analysis 22. Annex I – Rankings. 4.

(5) List of Tables. 1) Economic classes 1.1) % Class ABC 1.2) % Class E 1.3) % Class D 1.4) % Class C 1.5) % Class AB 2) Household Per Capita Income 3) Inequality index 4) Sources of Household Per capita income (average income) 4.1 ) Income from work 4.2 ) Private transfers 4.3 ) Public transfers – Family grant 4.4 ) social security benefits up to 1 minimum wage 1 SM 4.5 ) social security benefits above 1 minimum wage 1 SM 5) Participation of different incomes in total income 5.1) Income from work 5.2 ) Private transfers 5.3) Public transfers – Family grant 5.4 ) social security benefits up to 1 minimum wage 1 SM 5.5 ) social security benefits above 1 minimum wage 1 SM. 5.

(6) Preface The choice of Brazil and Rio de Janeiro, in particular, as host to the 2016 Olympic Games is justified by different reasons mentioned in the speeches of the Brazilian authorities in Copenhagen, namely: the opportunity given to the country and the SouthAmerican continent in general to join Olympic nations, the fact that Brazil will host the 2014 World Cup works as an intermediate objective that will enable the achievement of a greater challenge, that of organizing the Olympic games, and finally an perhaps more importantly, the marked improvement in Brazilian social and economic indicators in the last years, including the post-crisis period. More than the level of Brazilian development, its emerging power status would be the reason of our bid’s success. In this sense RIO 2016 is above all, a bet on Brazil future. Looking into the future, the Olympic challenge would be as a relevant incentive in the margin for Brazil and Rio de Janeiro city to focus on the achievement of tangible and more permanent improvements - the Olympic legacy - in the quality of life of its population. Similarly, the 2014 World Cup would place similar objectives to the attainment of social indicators for each one of its respective hosting cities. The results of various researches by the Center for Social Policies at the Getulio Vargas Foundation (CPS/IBRE/FGV see www.fgv.br/cpc/fc) including results based on the recent released 2008 PNAD microdata which shows that, between 2003 and 2008, nearly 20 million people left their poverty status in Brazil and more than 30 million people (equivalent to half of the French population) have risen to the new middle class (ABC classes) in five years, of which 6,7 million last year alone. This was at mentioned by President Lula in Copenhagen. In the current decade, the high income inequality that characterizes the Brazilian case has suffered successive decreases. For instance, while the 10% poorest Brazilians experienced real per capita increases of 72%, the 10% richest grew 11% in the period. The objective of this research is to describe the evolution of these income based social statistics across Brazil including municipalities, metropolis, states and regions. Where has Brazilian income grown more? That is the question. Was it in the Northeastern hinterlands or in São Paulo periphery? And the New Middle Class in Campo 6.

(7) Grande, has it already? What does explain the class changes in each place: growth or income redistribution? Why has inequality fallen so little within the poor Brazilian Northeast states while their mean income grew? Was it the impact of minimum wage? And the geography of poverty, did it change? And if it did, why? What is the capital of the Family Grant (Bolsa Familia)? And how about the pensioners’? Who is the champion of income and jobs generation? After providing the big Picture of Brazilian social indicators, the research explores the whereabouts of the level and changes in social indicators based on household per capita income.. As we did last year, we have extended our national-level. research to its regional (regions and states) and local specificities. Opening data for the capitals of the 27 states and the peripheries of the largest metropolises is an innovation of this new research. It will enable an evaluation of the mayors’ performance until their mandate’s last year, as well as changes brought about during the governors’ mid-term in each state, just as we did last year. Initially, the present research focuses on the social and economic performance of the 27 capital cities in Brazil, including the 2016 Olympic city and the 12 cities of the 2014 World Cup. We analyze the performance of the social indicators among the four last Olympic Games, More than a sports and social curiosity, there is an exact juxtaposition of the Olympic cycles with the mayors’ administrations. We present a recent portrait of the social indicators in these places for 2008, the last year of the mayors’ administration. The objective is to give transparency to the evolution of these indicators, holding municipal authorities accountable for the previous social performance. The similarity of the major sports competition in the world that the largest Brazilian cities will host in the next decade, the competition for indicators among these capitals may encourage the State and society to fight for a better society departing from tangible indicators. Besides local and national results, the research also analyzes international data, based on microdata from 128 countries. The Brazilian specific features in relation to their future: among 128 countries, we have the highest individual happiness perspectives (that is we are number 1 ) but the perception of the country’s overall happiness in the future falls behind (we are 43rd).. 7.

(8) Introduction. Macroeconomic analyses are aggregated in the very definition of the field of study founded by J. M. Keynes amidst the great depression in the 1930s. That is, looking to the group of people in a given society, not mattering their respective housing region, economic class, income sources composition, or other individual attributes (sex, age, etc.). It works as if it was a hybrid representative agent (half man, half woman) middle-aged, belonging to the middle class possibly from Minas Gerais (as political scientists say it, Minas Gerais summarizes the average Brazilian population’s diversity). In various situations, macroeconomic fiction proves adequate so that we do not get lost in unnecessary details, but in other situations, the devil lives exactly in the omitted details. In particular, in a country of continental dimensions, with huge inequality and, which becomes internationally known for the dissemination of old and new income policies such as Brazil, the aggregate analysis hides more than it reveals. Both in the period of expansion of the Brazilian income until September 2008, and in the period afterwards, there lacks a clear vision about at least three points: i) Who changed? Changes in the economy; who loses and who gains in terms of income classes (E, D, C and AB). The new Brazilian middle class became a crucial macroeconomic asset to compensate for the decrease in exports of our products as a result of the global retraction. The demand injection now is the key point today, but we are looking at the economy in a very aggregate fashion. To the extent that each segment has distinct expenditure trends, there are macroeconomic implications depending on who advances and who retreats in each segment. Without looking at the detail of the emerging and plunging groups, be it in the discovery of market niches, be it in terms of widening the social networks to losers. Ii) what has changed? Which income segment grew more or less, before or as a result of the crisis and the actions against it: work, Family Grant, pension or none of the above. This analysis of the close determinants of income will help us discover the reasons for change. It is not only about knowing what generates the new demand, but where the offer finds (or loses) this new (or old) demand, which leads us to the last and maybe more important dimension in this study, iii) where has it changed? In which regions, be they macro-regions, states, capitals, types of cities, did. 8.

(9) the domestic economy advance and where did it lapse? The present study addresses these three types of questions, using the last spatial questions as an axis to orient the society in the light of PNAD and PME microdata that were recently launched and see the geography of recent changes in income. Our questions: A recently launched research by the Center for Social Policies at the Getulio Vargas Foundation (CPS/IBRE/FGV) showed that 32 million people, or half of France, rose to class ABC between 2003 and 2008, of which 6,7 million last year alone. Where has Brazilian income grown more? That is the question. Was it in the Northeastern hinterlands or in São Paulo’s periphery? And the new middle class in Campo Grande – has it already proved its value? What explains class changes in each place: growth or income distribution? Why has inequality fallen so little in some Brazilian states? Impacts from minimum wage or a labor boom? And the geography of poverty, has it changed? Why? What is the capital of Bolsa Familia? And the capital of pensioners? What is the Champion of income and job generation? The present research explores the “whereabouts” of the level and changes of social indicators based on household per capita income. Just as we did last year, we extended in this second stage, our traditional national-level research to the regional (regions and states) and local contexts. Opening data for the capital cities of the 27 states and peripheries of the largest metropolises is an innovation of the present research. It will enable us to assess the mayors’ performances until the last year of their terms, just as changes brought about up until half the governors’ terms in each state, just we did last year. Besides drawing a map of the levels and variations of social indicators based on income until the end of 2008 based on PNAD, we have identified the evolution from pre to post-crisis periods of the income of different classes of Brazilian workers in the largest Brazilian metropolises, opening data for respective capitals and peripheries (i.e., the group of cities that are not capital). And we found out that Brazilian peripheries, much like peripheral countries, are thriving despite the crisis. The research maps within the Brazilian territory the evolution of social indicators based on household per capita income traditionally generated by CPS, such as 9.

(10) poverty, inequality and economic classes (AB, C, D, and E) summing up what happened to families from different economic classes in different locations. We analyzed the impact of different income sources. For instance, what was the relative importance of work earnings, social security benefits or Family Grant benefits in explaining the sources of changes in poverty, inequality and class sizes in each region, state, metropolis and capital city. Your answers: The research website: www.fgv.br/cps/atlas through interactive databanks will enable users to carry out a cross-reference analysis of information according to their particular interests in more detailed geographic levels than usual (e.g. regions, states, etc.) such as identifying state capitals and large urban centers’ peripheries. Where social indicators have grown more, was it in Rio de Janeiro recently elected the host to the 2016 Olympic Games or Sao Paulo? What does explain such changes in each place? And the geography of poverty, has it changed? If so, why? Who is the champion in the level and in improvements in social indicators among capitals in Brazil? Without extending this research further, it provides the ranking of the level and changes of social indicators in the 27 capitals.. 10.

(11) Rio 2016: a shock of progress? “Rio2016 represents a bridge between the means and the end, from the “illegal, so what?” to the “legal, what now?” overcoming obstacles between a shock of order and a shock of progress” Rio de Janeiro was a mistake. Let me, a true Carioca, try and explain. When the Portuguese arrived on the Brazilian shore in January, supposedly, they mistook the Guanabara Bay for a river estuary, and named it January River (Rio de Janeiro). Had this mistake not occurred, the Baia de Janeiro (January Bay) would be here. The Baia de Todos os Santos (All Saints Bay) inspired Bahia, the great rivers in the south and north of Brazil inspired the names of Rio Grande states. Here, the initial lapse was perpetuated in the name of the state, the capital city and Greater Rio, the metropolis, are composing what we call here three Rios. The mistake at the beginning of Rio’s history goes upstream: PDBG (Guanabara Bay Clean-up Program), financed by the Japanese Development Bank (our former Olympic rival). Besides money, there was society mobilization that gathered force in the Rio 92 Conference. Our research with the Trata Brasil Institute shows that the expansion of the general sewerage system in Rio came to a halt. Baia Azul, Salvador’s equivalent to the PDBG, financed by the IADB, doubled the access to basic sanitation infrastructure between the 1998 and 2002 World Cups. If money and mobilization are needed, so is good management. In the case of Three Rios, bad management, apart from an being an internal problem of their respective public authorities, is also present in their co-action. Octavio Amorim argues that when successive mayors and governors in Rio aimed to run for the presidency of Brazil, they stopped the flow of federal resources to the three Rios. Alignment between the three levels of government, which was clear in Copenhagen, is the exception – not the rule. The relationship between the State and society has also taken a step back here. While the country lived, in the past ten years, a labor/business formalization process, the Three Rios moved in the opposite direction, at least for the first part of this period. Between 1997 and 2003, all indicators regarding the formal status of small businesses, which characterize the local productive fabric, have fallen to half, just as did the level of workers’ social security adhesion in general. In this period, 11.

(12) the well-known local street-wisdom, the true local sport, had as its capital the neighbourhood of Lapa, and as its symbol a typical loser, not so much Disney’s character José Carioca. A hit with the public and the critics, O Globo newspaper’s series entitled “Ilegal, so what?”, illustrates the Carioca pledge against chaos, which reached the current governor and mayor and was implemented as actions called a “shock of order” in slums, streets and illegal constructions. The term “shock of management”, borrowed from the successful initiatives in the states of Minas Gerais and Espírito Santo – success both for its administrative trademark and the good performance of social, economic and electoral indicators, thus re-electing governors with 80% of local votes. This showed that a tangible target-oriented management may lead to a sea of results. The three Rios launched their respective shocks of order, moved on to shocks of management, but we are still in their means, not their ends; we have met the necessary, but not sufficient, conditions. There is where the success of Rio’s Olympic bid comes in, representing a bridge to link the “illegal, so what?” to the “legal, what next?” steps; to make this crossing between the two margins overcoming existing barriers between shocks of order and management, on one side, and the shock of progress, on the other side. We may now perhaps take advantage of Brasilia’s 50-year anniversary in 2010 to leave behind any nostalgic feelings about Rio’s status as the former Capital of the Republic, who remains attached to its past. Our sailors Amyr Klink and Torben Grael are living examples of the capacity of three Rio’s inhabitants to navigate towards new goals, as impossible as they may appear to be. The election of Christ, The Redeemer in 2007, a 75-year old pied as one of the new seven wonders, reflects this capacity. I was in Machu Picchu and I saw the Peruvian’s surprise about it, after all it was not a competition about our irrefutable natural endowments, but about human work. Peruvians may not know that our great human work at stake is the cohesion of Cariocas with the rest of the country around the presented objectives. Each Carnaval shows our renewed capacity to overcome challenges. Now, as many may be mistaken for a long time and be deluded by speeches about a future as bright as unlikely, CPS launches a research to monitor social indicators of the 27 Brazilian capitals. Our methodological innovation lies in opening microdata for the capital cities in PNAD, in order to compare mayors and their work. We begin by taking 12.

(13) advantage of the coincidence between mayors’ terms and Olympic cycles, comparing the performance of different mayors from the 1996 Atlanta Olympic Games to the 2008 Beijing Olympic Games. For instance, who was the champion in lifting people out of poverty: Cesar Maia II (Rio´s mayor between Sidney 2000 and Athens 2004) or Cesar Maia III? Where was the new middle class leap, quoted by Lula in Copenhagen, higher? In the small Campo Grande or in the Greater Sao Paulo area? What has changed in these cities? Why? Unemployment decrease, salary increase or none of the above? Obviously, different contexts must be considered, so as looking at different relative speeds among places, as in a healthy race for the best indicators in different periods of time. Beyond 2008 PNAD, we bring up data from the last twelve months when not only was the economic crisis occurring, but new mayors were taking over. More than a fixed route, the site WWW.fgv.br/cps/2016 is a navigation tool, allowing for a comparison of the capitals’ performance, including the 12 hosting cities of the 2014 World Cup. Winning the 2016 Olympics bid is just the start of the race, with hurdles, for the tangible results which is just beginning, and as any competition, it should be watched by the public. The Olympic Games, besides being of global interest, should produce a local legacy of the people, by the people and for the people in Brazil in general and of three Rios, in particular.. 13.

(14) General Brazilian Background The French associate the blue, red and white of their flag to the ideals of their revolution: freedom, equality and fraternity. What attributes would you choose to relate to our green and yellow flag colors? I would say diversity and inequality. Inequality is a Brazilian feature that has remained unscathed through the centuries. Yellow from the extracted gold, initially by native Indians, then by Africans - the last in the Western hemisphere to be freed. We live in a rich and unequal country. Our diversity is in Brazil and in each Brazilian, having been compared to a boiling mixing pot of ethnicities, creeds and behaviours. In our pseudo-democracy, everyone has the same colour, as I say it, various shades of green1. Green is the colour of the diversity in the mixed races – it is a secondary colour that results from the combination of yellow and blue, and would thus capture the Brazilian diversity. Diversity and inequality are our marks, which stop us from seeing the palette of Brazilian colours. At each PNAD, Brazilian society has the opportunity to look at the colours and faces as one stares at the mirror; to know the hindrances and advances of the year that stays behind. PNAD data reflect the answers of people about themselves, true self-portraits. PNAD’s expansion factors help to gauge the absolute size or the relative position of each social group among themselves and in relation to Brazil, keeping the original scale. Simultaneously to publicizing its rich collection of tables and analyses, IBGE releases the research’s microdata with an annual sample of more than 380 thousand individual answers to a good questionnaire with over 100 questions asked with exactly the same structure every year in the last two decades. Beyond its transparency, the flexibility and precision of the large microdata sample help to portray the relations between many facets of Brazilian life: school, work, pensions, etc.; girls, blacks etc. For instance, as the school impacted on work or how work impacted on the income of families, etc.. 1. Only that in Brazil the darker shades of green usually live in slums and can only get through the rear entrance of buildings inhabited by lighter shades of green. In France, diversity is a concern of a different nature, it is not uncommon to meet French people who say “No! Vive la France, I want to remain apart, to keep my culture”. Brazilian diversity’s green is within each one and not stuck in groups or primary colours.. 14.

(15) Without entering virtually infinite possibilities to cross microdata, the question that remains in each PNAD is what we can conclude from the tables. Researchers, managers, journalists and ordinary citizens – all of us – are drowned in numbers! The objective of this study is to build a brief profile of the Brazilian conditions from a range of PNAD information. The literature on social well-being seeks to synthesize the various aspects of reality for different people. The chapter on social indicators based on income translates data on salary, journey, occupation, unemployment, pensions, access to social programs, etc. into fewer numbers, each one with the capacity to portray a peculiar aspect of life in society, such as well-being level, inequality, poverty rate, economic class composition. A first effort is to condense information in order to transform it into practical knowledge about how much the Brazilians’ wealth has grown or decreased. We recognize that the exercise is a simplification of reality whereby the richness of information and PNAD’s colourful possibilities become black and white paintings. Roughly, it is worse than attaching values to certain artworks, as here we are talking about the lives of people – in the case, Brazilians.. Evolution Panorama: Social Measures based on Per Capita income Given its national and annual nature, PNAD enables the monitoring of the evolution of various social indicators based on income. The release of PNAD data is the moment when society faces a sort of mirror, sees its face and assess its own advances and obstacles. The panorama available in the research website presents a time evolution of different indicators such as poverty, economic classes, income, inequality and education (among other indicators), since the beginning of the 1990s. There follows the group of variables available for analysis. Each one of these indicators may be analyzed for the overall population or by open sub-groups: i) sociodemographic features such as sex, age, years of schooling, race, position in the household; ii) characteristics of the producer, including job position, social security tax, education and access to digital assets; iii) characteristics of the consumer such as access to consumer goods and services; iv) spatial features such as housing, area (metropolitan, non-metropolitan urban and rural), states and, innovatively, capitals and peripheries:. 15.

(16) http://www.fgv.br/ibrecps/RET4/CPC_evolucao_temporal/index-eng.htm. Income distribution. We present next the accumulated income gain per decile of population, starting with the poorest groups between 2001 and 2008. The growth rate decreases as we move from the first (72,2%) to the last decile (11,1%). Although the growth recovery period only began after the end of the 2003 recession, inequality had fallen beforehand.. Accumulated Variation of Average Income by tenth of income - Brazil (2008/2001) 72,45% 55,50%. 50,76%. 45,39%. 40,52% 37,46% 31,65% 25,74% 19,18% 11,37%. 10. 20. 30. 40. 50. 60. 16. 70. 80. 90. 100.

(17) The redistributive movement in terms of the Gini index is translated in the graph below. Gini is the most popular inequality measure applied to the most disseminated concept of well-being in the literature, which is household per capita income, including null incomes.. Gini Index. Source: CPS/FGV based on PNAD/IBGE microdata. Gini-measured inequality drops to 0,5486 in 2008, a decrease of -1,15% reaching a similar speed to that observed in the decade of inequality reduction:. 17.

(18) Poverty, Growth and Inequality Scenarios The proportion of poor people in Brazil, according to our estimates on the 2008 poverty line is 16,02%. Initially, in a neutral scenario in terms of distribution, if national per capita income grows 2,6% per year in the next seven years, which would correspond to 4% a year of total income growth for the overall population growth, poverty would drop to 12,43%, a decrease of 22,38%. This reduction would be even higher if the mentioned accumulated expansion of 20% was combined with a decrease in the Gini index, as it was similarly observed in the last seven years. This would equate the income distribution of Rio Grande do Sul in 2007. In this case, poverty would drop 41,22%, leading to a poverty rate below the two-digit mark: 9,45% of the Brazilian population would be considered extremely poor. However, even with a null per capita income growth, if inequality fell to Rio Grande do Sul levels, gains would still be reasonable – for the poor, of course – poverty rate would drop 18,58% in the period, reaching 13,04% of the population. This decrease in poverty is enough to overcome the pace of the first Millennium Development Goal of 50% decrease in 25 years, dropping 52% in this period. This note illustrates the prospective and retrospective role played by the reduction in income inequality within the Brazilian context.. Reasons for Change: Inequality. How do we reduce inequality? Once more, the present decade can show us the way by applying the decomposition methodology of the Gini variation2 to the period 2001-2008. Income from work explains 66,86% of the inequality reduction, then come the social programs, in particular the Family Grant and its precedent School Grant, which explain 17% of the inequality reduction, while the social security benefits explain 15,72% of the income de-concentration; remaining incomes account for less than 1%. It is desirable that the analysis considers not only the impacts of different income sources, particularly transfers from the Brazilian state, on inequality displacements, and the public resources as well.. 2. Hoffman 2006 and Soares 2006 apply this methodology to Brazilian data at the beginning of the decade. Kakwani, Neri and Son 2005 and Barros et all. 2006 apply other methodologies to the same data.. 18.

(19) Reasons for Change: Mean Income If something changed, the second struggle is to find out: why has it changed? How has it changed? These last questions suggest the two complementary lines of answers explored here, knowingly: the first one looks at the approximate determinants of the income distribution and the primary components of people’s income, the role of pensions, social programs and work (and its components) in the various synthetic measures. Decomposition of Income into different sources PNAD. Year. Income from all sources. Income from all jobs. Other private incomes. Public Transfers BF*. Social Security minimum MW*. Social Security above > MW*. 2008 – R$. 592,12. 450,29. 12,86. 12,73. 28,05. 88,2. 2008 – % Composition. 100%. 2,15%. 4,74%. 14,90%. Annual growth rate 2003-08. 5,26%. 5,13%. 2,62%. 20,99%. 6,64%. 4,44%. Growth 2007-08. 5,49%. 4,5%. 15,13%. 30,83%. 1,63%. 7,68%. 76,05%. 2,17%. Source: CPS/FGV based on PNAD/ IBGE microdata.. Between 2003 and 2008, the average per capita income of Brazilians increased 5,26% in real terms (i.e., population growth and inflation have been discounted) going from 458 to 592 reais per month. Income sources that increased more were social programs (20,99%) influenced by the expansion of the Family Grant, created in 2003. Next, came the share of pension income attached to the minimum wage (6,64%). The effects of minimum wage readjustments, as it increased more than 45% in the period, put a pressure on the basic value of benefits and the number of elderly, as a result of the population’s aging process. Income from social security above the minimum amount grows less than the general income growth. It is worth pointing out that income from work has had an average increase of 5,13% per year, which grants a sustainable support for the living conditions beyond the official income transfers. Income from work corresponds to 76% of the average income perceived by the Brazilians and 75% of the income gain observed has come from there. In the last year, the growth of per capita income from work and pensions bound with the minimum wage is a little lower, and social programs reach 30,8%. In any case, in both periods – although there has been a strong increase in income from social. 19.

(20) programs and pensions tied to the minimum wage – the share of income from work remains close to the impressive growth in income in this phase of boom.. Brazilian Income The main feature of the approach used here is its level of disaggregation into four groups of income.. Class AB: according to the last PNAD, the share of people in class AB (household income above 4807 reais) grew 7% in the last year, which corresponds to 1.5 million entering the upper income group. In the last 5 years, 6 million people have ascended to this class that, in 2008, reaches 19,4 million people.. Class C: The same class that reached 37,56% of the Brazilian population in 2003, reached 49,22% in 2008, or 91 million people with an income between 1115 and 4807 reais monthly, the dominating class in terms of population size. This accumulated growth of 31% in 5 years means in population terms 25,9 million Brazilian people who had not been and became class C in the last 5 years (5,3 million last year alone).. Class D: the proportion of people in class D is 24, 35% in 2008 reaching 45,3 million Brazilians with an income between 768 reais up to the lower limits of class C. In terms of trends, there has been a reduction of 0,9 million in one year or 3%, and 1.5 million if we consider the last five years.. Class E: With a reduction of 12,27% in the last year, or the exit of 3,8 million people out of the lowest income group up to 768 reais monthly, the poverty level in our methodology. This movement corroborates a trend since the end of the 2003 recession, when poverty fell 43%, that is, around 19,4 million people have crossed the poverty line. As a result, 29,9 million poor people (16,02% of the population3) who would be instead 50 million had poverty not fallen in the last years.. 3. With an income below 137 reais monthly (greater São Paulo area prices or 145 reais at national average prices pondered by the population of each class).. 20.

(21) Composition of Income per Economic class In the period from 2003 to 2008, we noticed that the share of income associated with social programs, such as the Family Grant, doubled. This corresponds to the poor groups by the national average line of CPS – after the increases announced by the government and the new entry criterion for the Family Grant, the share of these programs in the respective incomes increased from 4,9% to 16,3%. The analysis of the participation of different income types per economic class may be useful to assess the prospective impacts of different public policy tools on income distribution, such as for example the measures adopted with the external crisis context in September 2008, namely: Increases in the Family Grant and other programs not related to the social security tend to benefit predominantly class E that has 16,25% of its earnings from this type of income.. It is interesting to separate income from social security benefits as individual earnings up to one minimum wage and benefits above this minimum, because distinguishing among such increases was stressed in 1998. The major beneficiary of the increase in the social security minimum (basic) benefit is class D, with 12,66% of income tied to it. Finally, the increase in pensions above this minimum value benefits above all class AB as 18,94% of its earnings derive from this source. This measure is being discussed today.. Source: CPS/FGV based on PNAD/IBGE microdata. 21.

(22) Social Draw Nine months after the crisis started, there is an already clearer vision about its effects in the Brazilian people’s income in the six major metropolitan areas in the country. Income inequality underwent a serious deterioration in January, when part of last years’ gains were lost, but it has now come back closer to its pre-crisis levels. Even class AB, that earns more than 4800 reais per month and who had lost more with the crisis (2,7% in January alone), is today only 0,5% below one year ago’s levels (14,97% of the population is in class AB with 55% of the country’s income). Class C is already in a positive situation with a 2,5% gain in 12 months (53,2% it is the dominating class in terms of population size). If this draw may be considered a good result in times of crisis, on the other hand, it hides a sudden halt of the previous improvements in the indicators: from July 2003 to July 2008, class AB grew 35,7%, class C increased 23,1% and income inequality dropped as it had never dropped before in the Brazilian statistical series. Looking on the bright side, 2008 PNAD must provide, despite the current crisis, a more or less faithful portrait of the living conditions in 2009.. Variation of Economic Classes Pre versus Post-Crisis Classe AB. Classe C. Classe D. Classe E. 23,1%. 2,5% -0,5%. -4,1%. -3,3%. -15,5%. -37,0% jul03 a jul08. jul08 a jul09. Source: CPS/FGV based on PME/IBGE microdata. 22.

(23) Strategy to widen the scope of economic classes Our strategy is, at each update of our traditional series based on household per capita income such as poverty, inequality, social welfare and now income classes, to include a new dimension to the analysis of the various economic classes: entrepreneurship. (www.fgv.br/cps/crediamigo2),. (www.fgv.br/cps/crediamigo3),. micro-insurance. microcredit (www.fgv.br/cps/ms).. (www.fgv.br/cps/crediamigo2), exploring in each research a new perspective. In the current research we will explore a multidimensional view from the rich data offered by PNAD. Still exploring the rich microdata of the PNAD/IBGE we applied a model of sequential variable selection according to the level of statistical significance related to household per capita income, always based on the household/family as the basic unit. We look to the producer and consumer. In the case of the consumer, a range of information on access to consumer goods, housing and public services are provided by PNAD. In the case of the producer, the focus is on the inclusion in the labor market that reflects human physical and social capitals, not only the education level of the household’s person of reference and his/her spouse, but also the investment on the future of their kids open by age groups and types of school.. What are the main stocks associated with income flows?. i. Technical aspects We present initially in this subsection a discussion on the the series of estimating models Apresentamos inicialmente nesta subseção uma discussão sobre uma série de modelos de estimação dos determinantes das classes econômicas. Multivariate Analysis – Methodology The bivariate analysis captures the role played by each attribute considered isolatedly in the demand for insurance. That is, we desconsider possible and probable interrelations of the explanatory variables. For example, in the calculation of insurance by state within the Federation, we don’t consider the fact that Sao Paulo is a richer place than most states, thus should have greater access to insurance. The multivariate analysis used further ahead seeks to consider these interrelations through a regression of the many explanatory variables taken together. Aiming to provide a better controlled experiment than the bivariate analysis, the objective is to capture the pattern of partial correlations between the variables, interest and explanatory. In other words, we have captured the relations between the two variables, keeping the remaining variables constant. This analysis is very useful to identify the repressed or potential demand as we compared them, for instance, which are. 23.

(24) the chances of a person with more education having higher income, if he/she has the same characteristics as the comparison group.. Choice Models for Explanatory Variables We being exploring the wide range of information relating to the possession and use of assets based on PNAD, using a model of variable selection according to the level of statistical significance to explain the household per capita income. It is worth noting that both in the field of traditional social indicators (i.e. poverty and welfare4) just as in the definition of economic classes (i.e. E, D, C and AB) the family/househodl is the basic unit of analysis under the hypothesis of its members’ solidarity who, on the whole, share the earnings much like the “all for one and one for all” of Dumas’ Three Musketeers. Another point is the use of income as a unit of reference to integrate different information on access and use of productive and consumption assets. In our view, whether people like it or not, income is the most used variable in economics and, if we want to increase the dimension of the analysis, it is interesting to use what has already been done in practice. Here it is important to note that we speak of the sum of different income sources reported by people to PNAD and not the aggregate vision of production implied in the GDP5. Afterwards, based on the selection of variables, we tracked variables referring to the producer and consumer available on PNAD. The exercise is part of a learning process to decide what matters in the definition of classes and how much each of the estimated components are worth. In order to determine which of them have higher explanatory power and which will be more relevant, by applying a variables’ sequenced choice procedure that uses a mincerian income equation model. The list of selected variables for each model (from a test F) is provided next, in increasing order of importance, in a self-explanatory list of 31 groups of variables; the eliminated variables are not reported in the table:. 1. ORDEM DE ENTRADA NO MODELO Number of per capita toilettes. 4. Welfrae as inequality measure derived from the social welfare measure used. the Stiglitz-Sem report, launched on sept 15 de setembro, defends the use of information from household surveys and less information based on GDP (per capita PPP) that prevails in the analysis. This is described in the introduction. 5. 24.

(25) 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29. Telephone Spouse’s education Type of family Head pays social security tax Washing machine Number of per capita bedrooms Head’s education Spouse’s job status Children in school(7to 14) Children in school (0 to 6 y.o.) Head’s job status Computer Fridge Children in school (15 to 17) Type of household (owned, financed, rent) Head member of union Freezer Per capita number of rooms Sewerage system Radio Number of toilettes Number of members Television Waste collection Age when head began to work Number of rooms Income from work participation Number of bedrooms. 25.

(26) Initially, it is worth noting that we purposefully omitted sociodemographic and spatial variables from the explanation of per capita income so that we could infer afterwards which is the equivalent income of people with different features. We should also mention that the variable number of toilettes, followed by access to mobile telephones well before completed years of schooling of the reference person comes in 8th place (3rd place in the case of the spouse education) that typically has the highest explanatory power in empirical researches on income inequality and poverty. Obviously, we are not attempting to establish a causal relation between different variables of stock and income flow, even more because this is a two-way relation. In our interpretation, we will identify variable that are more dependent on income that incomegenerating. The exercise helps to gauge the structure of the model that assign equivalent income and its counterparts in terms of consumption potential and income-generating capacity. The table is self-explanatory.. 26.

(27) Income simulator: Total, consumer and producer Tool used to simulate total income of the population through a combination of individual attributes tied to the consumer and producer. For that, you must select characteristics in the form below and click on simulate.. The graphs show the total household income, in the following order: - Total: expenditure potential and income generation - Consumer perspective: expenditure potential - Producer’s perspective: income generation. One of the bars represents the current scenario, according to the selected attributes; the other is the previous scenario as simulated before.. http://www.fgv.br/ibrecps/cpc/SIM_PNAD_0208_RENDATOT/renda.htm. 27.

(28) In the table above, we are not considering the magnitude of each category’s coefficient, but the power of the categories taken together to explain income. Wen looking at the magnitude of extreme coefficients in each variable, the equivalent income of a person who lives in a household with one bathroom for each person, if we double it (two bathrooms for four people instead of one bathroom) the income increases 27,5% in relation to the previous scenario; while a person with a fixed landline and a mobile at home, ceteris paribus, and a person with none of these communication technologies at home, it is 38,5% higher. The spouse variable is more significant than the reference person, as it refers not only to the impact of education, but to the composition of family income which is more or less diversified by virtue of potential income earners. The same does not happen to the education variable of the reference person, because each household has at least one reference person. A household with a spouse with at least incomplete university degree (12 or more years of schooling) has a 28% higher income than a spouse with an ignored educational level (regression basis, zero coefficient), which in turn has an income 14% higher than those without spouse. Following the order of statistical relevance of the variable selection model, we have a variable on the type of family composition where a family consisting of a couple with all children under 14 has a per capita income around 30% lower than a family with a couple and no kids. The fifth variable with higher predictive power is that which captures the nature of the public or private social insurance of the household head or none of the above, that is, inactive or unemployed household heads, where the reference person pays both types of insurance with a family/household per capita income around 30% higher than those inactive or unemployed. The remaining variables may be analyzed through the same prism or through the simulator described in the box further down. In the annex, we present a conceptual explanation and a practical application of a nother methodology to define class based on principal components, and not based on people’s income. In essence, the principal component analysis reduces the variables for a group of artificial variables, which is accomplished through turning redundant variables into new variables that may be used in subsequent analysis as forecasting variables in a multiple regression – or in another type of regression. Technically, a principal component may be identified as a linear combination of optimally weighed variables.. 28.

(29) The first component extracted from a principal component analysis would be responsible for a maximum amount of total variance of the observable variables. The second component extracted would account for a maximum variance of the group of data that does not derive from the first component. In ideal conditions, this means that the second component would be correlated with some of the observable variables that do not show strong relations with the first component. The final practical result would have three components: the first and most significant could be interpreted as the consumer’s side – consumption of consumer goods’ variables (washing machine, fridge and freezer), and the absolute size of household captured by the number of rooms and toilettes. The vector captures two items that we also linked to the production as education of the household head and his/her spouse and the possession of goods tied to ICT such as fixed landline and mobile phones and computer connected to the internet. The second vector with more relevance may be defined as that consisting of some variables of quality on the household’s housing quality captured based on the per capita number of room, bedrooms and toilettes, the number of members in the household and the presence of teenagers as well as variables of education quality of the children between 7 to 14 years old and 15 to 17 years old, as well as the type of family structure (family headed by a mother with children of up to 14 years old). Finally, the third factor may be interpreted as effective generation of income consisting of the participation of income from work in the total income, the type of social security/insurance (public and private, etc.) and job status (private employee, unemployed self-employed, etc) of the household head.. Marginal contribution of the stocks to the inequality of flows We explored then the contribution of each variable of stock on the variance ofthe household per capita income inequality. We calculated the marginal contribution of each variable on the total R2 of the regression taking them one by one out of the complete regression and calculating the relative difference such as its contribution to the margin for the income inequality:. 29.

(30) Marginal Contribution of Income Inequality Without the respective variable R2. All variables (R2 original) Telephone Spouse’s job position Children’s school attendance(7toa 14 years old) Washing machine Education of the head of the household Type of family Head’s job position Computer Children’s school attendance(0toa 6 years old) Type of household (owned, financed and rent) Children’s school attendance(15toa 17 years old) Fridge freezer Spouse’s education Head pays social security tax Head belongs to union per capita number of toilettes per capita number of bedrooms per capita number of rooms Sewerage system radio Number of members Television set Age when head started working number of rooms number of toilettes Waste collection number of bedrooms Participation of work income. 30. dif R2 0,6924 -. dif R2/R2 orig%. 0,6813 0,0111 0,6825 0,0099. 1,60 1,43 0,92 0,81 0,78 0,77 0,72 0,69 0,58 0,52 0,49 0,46 0,40 0,39 0,38 0,12 0,07 0,06 0,04 0,04 0,04 0,03 0,03 0,01 0,01 0,01 0,01 0,00 0,00. 0,6860 0,0064 0,6868 0,0056 0,6870 0,0054 0,6871 0,0053 0,6874 0,0050 0,6876 0,0048 0,6884 0,0040 0,6888 0,0036 0,6890 0,0034 0,6892 0,0032 0,6896 0,0028 0,6897 0,0027 0,6898 0,0026 0,6916 0,0008 0,6919 0,0005 0,6920 0,0004 0,6921 0,0003 0,6921 0,0003 0,6921 0,0003 0,6922 0,0002 0,6922 0,0002 0,6923 0,0001 0,6923 0,0001 0,6923 0,0001 0,6923 0,0001 0,6924 0,0000 0,6924 0,0000.

(31) Mincerian Equation The mincerian equation of salary is the basis of a vast empirical literature of labour economics. The salary model by Jacob Mincer (1974) is the framework used to estimate the returns on education, among other variables that determine the salary. Mincer conceived an equation for earnings that would be dependent on explanatory factors related to the academic level and experience, besides possibly other attributes, like sex, for instance This equation is the basis of the labor economics particularly in what concerns the effects of education. Its estimate has already encouraged hundreds of studies, that tried to include different educational costs, such as taxes, fees, opportunity costs, learning material, just as the uncertainty and expectation of agents present in the decisions, the technological progress, non-linearity in school, etc. identifying the costs of education and work earnings enable a calculation of the internal rate of return on education, which is the discount rate that must be compared to the market’s interest rates to determine the optimal quantity of investment in human capital. The MIncer equation is also used to analyze the relation between growth and educational level in a given society, besides inequality determinants. The typical econometric model of regression of the mincerian equation is: ln w = β0 + β1 educ + β2 exp + β3 exp² + γ′ x + є where w is the salary earned by the individual; educ is its educational level, measured by years of schooling; exp is its experience, whose Proxy is the individual’s age. x is a vector of the observable characteristics of the individual, such as race, gender and region; є is a stochastic mistake . This is a model of regression in the log-level format, that is, the dependent variable - salary – is in a logarhytm format and the independent variable, more relevant – education – is on level. Therefore, the β1 coefficient measures how much an extra year of scholing causes a proportional variation in the individual’s salary. For instance, if β1 is estimated in 0,18, this means that each extra year of study will be related on average with na increase in salary of 18%. Mathematically, we have: Deriving , we found that ( ∂ ln w / ∂ educ ) = β1 On the other hand, by virtue of chain, we have: (∂ ln w / ∂ educ) = (∂ w / ∂ educ) (1 / w) = (∂ w / ∂ educ) / w) Logo, β1 = (∂ w / ∂ educ) / w, corresponding thus to the percentage variation of the salary from each unit increase per year of study.. 31.

(32) Decomposition of Inequality How do we reduce inequality? Once more, the present decade may show us the way, by applying to the period 2001-2008 the methodology of decomposition of the gini variations. According to the last column of the table below shows, the work income explains 66,86% of the inequality reduction expected between 2001 and 2008, next come the social programs with emphasis on the Family Grant and its predecessor School Grant, that explain 17% of the inequality reduction while social security benefits explain 15,72% of the deconcentration of income, and the remaining income account for a residue under 1%. The remaining columns of the tables compare 2008 to the other years. The following tables decompose the nature of the effects per type of income separating the effects of the contribution of each source to the total income, by the change in the relative mass of benefits from the effects of inequality of each source, based on the Total Income Gini. Percentage effect of each share of earnings in the change of the GIni index on the distribution of the household per capita income in Brazil Percentage effect Share Work earnings Social security Family Grant Private transfer Bonus Total Delta GINI. 2007 a 2008 116,15 -1,06 -1,03 -14,21 0,23 100,07 -0,0064. 2006 a 2008 89,30 21,67 -10,01 -0,82 -0,11 100,03 -0,0137. 2005 a 2008 64,07 25,49 11,43 -0,92 -0,04 100,03 -0,0196. 2004 a 2008 65,84 27,68 6,99 -0,40 -0,08 100,04 -0,0225. 2003 a 2008 66,39 17,56 16,83 -0,71 -0,07 100,00 -0,0344. 2002 a 2008 66,86 13,80 18,47 0,93 -0,04 100,02 -0,0400. 2001 a 2008 66,86 15,72 17,00 0,50 -0,05 100,02 -0,0471. 2007 a 2008 0,41 -0,20 26,45 -4,10 0,10 22,65. 2006 a 2008 -0,06 -0,06 -0,72 3,05 -0,10 2,11. 2005 a 2008 0,00 0,18 6,62 2,58 -0,09 9,29. 2004 a 2008 0,00 0,07 9,44 1,95 -0,05 11,42. 2003 a 2008 0,00 0,10 9,29 1,16 0,06 10,61. 2002 a 2008 -0,02 -0,18 4,71 1,75 -0,06 6,21. 2001 a 2008 -0,04 -0,21 5,28 1,32 0,01 6,36. 2007 a 2008 115,74 -0,85 -27,48 -10,12 0,13 77,42. 2006 a 2008 89,36 21,73 -9,30 -3,87 0,00 97,92. 2005 a 2008 64,07 25,31 4,81 -3,50 0,05 90,74. 2004 a 2008 65,84 27,61 -2,45 -2,35 -0,03 88,62. 2003 a 2008 66,39 17,46 7,54 -1,87 -0,13 89,39. 2002 a 2008 66,88 13,98 13,75 -0,82 0,01 93,81. 2001 a 2008 66,90 15,93 11,72 -0,82 -0,07 93,66. Composition effect Share Work earnings Social security Family Grant Private transfer Bonus Total Delta GINI Efeito concentração Share Work earnings Social security Family Grant Private transfer Bonus Total Delta GINI. 32.

(33) It is interesting that the analysis considers not only the impacts of different income sources, particularly the transfers from the Brazilian government, on the inequality movements, but also its costs to the public accounts. Costs of Poverty Eradication. Another useful measure in the design of public policy is the income gap (P1). That is, how much income lacks, on average, to the poor so that they can meet their basic needs in the market. Using as base our income insufficiency line, the average deficit expressed in monetary terms of each poor Brazilian would be R$ 56,29 monthly at average prices in Brazil. In 2007, the same statistics was R$ 57,44.Captured by the decrease in the index known as P2 (falls from 5,08 to 4,33) we observe a relief in the seriousness of poverty (in a counterpoint to what occurred last year). Data on income increase in the first decile and the decrease in the proportion of people with very low income measured in the poverty line indicate that the poorest of the poor gained in the period. Going back to the calculation of the eradication costs in 2008, as just part of the Brazilian population is below the line, data shows that it would be necessary R$ 9,01 on average, per person versus R$ 9,65 in 2007), to relief poverty totally in Brazil, a total cost of R$ 1,7 billion reais per month and R$ 20,2 per year. Information reveals how much it would cost to complete the income of each Brazilian up to the line of R$ 132 nationally or 145 reais at SP prices, that is, the lowest value of sufficient transfers to lift each poor person to the floor of their basic needs. This exercise should not be read as a defense of specific policies, but as a reference to the social opportunity cost of adopting unfocused policies. Data is useful to indicate the target of policies and organize their sources of finance.. Cost of Poverty Eradication. Minimum transfers to eradicate poverty. 33.

(34) Brazil 2008. R$ person 9,01. R$ total month. R$ total year. 1.680.719.363. 20.168.632.359. R$ non poor 10,73. R$ poor 56,29. Cost of Poverty Eradication. Wealth transfers per non-poor person. 0.5 % a.m. Brazil 2008. 1 % a.m.. 2147. 1073. Lorenz Curve - Brazil 2008 Household per capita income inequality. Source: CPS/FGV based on PNAD/ IBGE microdata.. Participation in Total Income 2008 – Brazil. 34. 2 % a.m. 537.

(35) Level by income groups 2008 – Brazil. Source: CPS/FGV based on PNAD/ IBGE microdata.. 35.

(36) II Mapping Brazilian Social Progress (income based measures) Objective The objective of this study is to provide a short account of the different income earned by Brazilians, seeking to sum up the various aspects of different people’s reality. The chapter of indicators based on income in the social welfare literature translated the data of salary, working day, occupation, unemployment, pensions and benefits, access to social programs, etc. into few numbers, each with the capacity to depict a peculiar aspect of life in society, such as welfare level, inequality, poverty rates, and economic class composition. A first effort is to condense information to turn it into practical knowledge, such as how much income in Brazil has grown or diminished in different places. The second effort is, once measurement is defined for the whole, to tread the inverse way opening per capita income for large income types in order to understand the close determinants of poverty. In all cases, the core of the research will focus on the spatial opening of income information.. The Geography of Poverty Between 2003 and 2008, there was a 43,03% reduction in poverty, corresponding to 19,3 million people leaving poverty defined as income below 137 reais in household per capita income. Illustratively, we will initially work with a more local opening. Among the 27 capitals of the Brazilian states and the peripheries of the six major metropolises, between 2003 and 2008, the city of Palmas stands out as the highest rate of poverty reduction (-80,9%) and, among the lowest reduction rates are the city of Rio de Janeiro (-34,8%) and Recife´s periphery (-36,4%). In terms of the series’ levels in 2008, the lowest poverty rates are found in Florianópolis (2,36% of the population) and Curitiba (3,92% and the highest are in Maceio (25,6%) and once more Recife’s periphery (26,4%). In the table below we present the five top and the worst five cities in terms of poverty rates in 2008, as well as their position in previous years and in variation rankings. The complete rankings with all 36 places may be found in annex 1.2. 36.

(37) % Classe E % rank 2008. 2008. % rank. 2007. % rank. Var (%). 2003. rank. Var (%). 2007/2008 rank. 2003/2008. 1 2 3 4 5. Periferia de Recife - PE Maceió - AL Periferia de Salvador - BA Periferia de Fortaleza - CE Recife - PE. 26.38 25.60 25.22 24.63 20.75. 2 7 5 1 3. 26.75 21.46 22.01 27.07 22.60. 5 4 1 2 6. 41.47 41.70 47.69 46.69 35.85. 10 5 6 17 15. -1.38% 19.29% 14.58% -9.01% -8.19%. 2 6 12 13 7. -36.39% -38.61% -47.12% -47.25% -42.12%. 1 2 3 4 5. Florianópolis - SC Curitiba - PR Goiânia - GO Vitória - ES Palmas - TO. 2.36 3.92 4.50 5.45 5.68. 36 34 32 35 21. 1.68 3.20 6.19 2.77 13.51. 36 35 32 33 17. 6.49 10.50 13.49 11.99 29.78. 3 4 28 1 36. 40.48% 22.50% -27.30% 96.75% -57.96%. 33 31 34 25 36. -63.64% -62.67% -66.64% -54.55% -80.93%. Source: CPS/FGV based on PNAD microdata (acesse anexo 1.2 ou http://www.fgv.br/ibrecps/RET4/CPC_evolucao_espacial/). Capital cities had 11,28% of their population living in poverty in 2008 versus 12,37% in the peripheries; and in 2003, right after the so-called metropolitan crisis, they were very close to one another with slightly higher poverty rates in capitals, 22,47%, versus 22,06% in the peripheries. . % Classe E. Total Capital Periferia das metrópoles (não capital) Urban area - non-metropolitan Rural area. %. %. %. Var (%). Var (%). 2008 16.02 11.28. 2007 18.26 13.77. 2003 28.12 22.47. 2007/2008 -12.27% -18.08%. 2003/2008 -43.03% -49.80%. 12.37 14.02 34.82. 13.87 16.09 37.30. 22.06 25.45 51.45. -10.81% -12.87% -6.65%. -43.93% -44.91% -32.32%. Source: CPS/FGV based on PNAD microdata (acesse anexo 1.2 ou http://www.fgv.br/ibrecps/RET4/CPC_evolucao_espacial/). Next, we open data for states, and the poorest are Alagoas (38,76%) followed by Maranhão (33,75%). Just as in the case of capitals ranking, the southern states present the lowest poverty rates, amongst which Santa Catarina has the lowest one with 4,53%, followed by Parana.. 37.

(38) % Classe E % rank 2008. 2008. % rank. %. 2007. rank. 2003. Var (%) rank. Var (%). 2007/2008 rank. 2003/2008. 1 2 3 4 5. Alagoas Maranhão Piauí Paraíba Sergipe. 38.76 33.75 32.38 29.20 26.56. 2 1 3 4 6. 37.93 38.30 37.05 33.19 28.59. 1 2 3 4 6. 57.66 55.68 52.01 47.28 41.58. 3 14 16 15 8. 2.19% -11.88% -12.60% -12.02% -7.10%. 1 8 5 6 2. -32.78% -39.39% -37.74% -38.24% -36.12%. 1 2 3 4 5. Santa Catarina Paraná São Paulo Rio Grande do Sul Minas Gerais. 4.53 6.13 8.79 9.01 9.27. 27 26 22 23 25. 3.67 4.50 10.86 10.03 9.76. 27 26 23 25 22. 8.29 14.08 17.65 14.24 17.67. 2 1 23 12 6. 23.43% 36.22% -19.06% -10.17% -5.02%. 13 25 18 4 16. -45.36% -56.46% -50.20% -36.73% -47.54%. Source: CPS/FGV based on PNAD microdata (acesse anexo 1.2 ou http://www.fgv.br/ibrecps/RET4/CPC_evolucao_espacial/). Complementing the poverty analysis, we now assess what happened to the 5 Brazilian macro-regions (excluding the rural area of the northeast). We found the highest poverty rates in the Northeastern region, 30,69% in 2008. Even though it does not present the highest decreases in the poverty rate, it is important to look at the absolute levels to capture the size of its reduction, as in 2003 49,81% of its population lived in poverty.. % Classe E. North Northeast Southeast South Center. %. %. %. Var (%). Var (%). 2008. 2007. 2003. 2007/2008. 2003/2008. 19.07 30.69 9.68 7.29 10.49. 22.37 34.20 11.60 8.03 11.78. 35.92 49.81 18.40 13.77 23.22. -14.75% -10.26% -16.55% -9.22% -10.95%. -46.91% -38.39% -47.39% -47.06% -54.82%. Source: CPS/FGV based on PNAD microdata (acesse anexo 1.2 ou http://www.fgv.br/ibrecps/RET4/CPC_evolucao_espacial/). As year I A.C. (after the crisis) was completed on September 15th, when the crisis occurred in the stock markets abroad, what can we say about its effect on the income of the Brazilian poor groups? We follow up here, with data up to July 2009, our monitoring of the evolution of the population composition in terms of different economic classes. PME enables a look into these types of areas in the post-crisis period (PME surveys the work earnings within the six major metropolises only). Comparing July 2008 with July 2009, labor poverty decreased more in the periphery of Belo. 38.

(39) Horizonte (-26,13%) and increased more in São Paulo capital (11,1%). Next, we present the poverty variation in this last period, which will be detailed further.. Variation in Metropolitan Poverty (Post-Crisis) – July 2008 to July 2009. Fonte: CPS/FGV a partir dos microdados da PME/IBGE. Real Poverty Decreases (and Class E) The CPS has been the first to release many results. The group that founded CPS was the first to show in February 1996 the improvement in social indicators after the Real Plan. In 1999, the group showed the increase in poverty in view of external crises. In 2004, CPS not only showed social deterioration in Lula’s term first year (2003), as well as the decrease in poverty in 2002 as the Cardoso administration ended. No other institution dared to launch a research on this theme. Access it on the website or through this link trajetória do CPS no estudo da pobreza. Looking at the large aspects of the poverty series since 1992, when the new PNAD questionnaire was implemented, we have two important changes in levels. Firstly, in 1993-1995, the proportion of people below the poverty line goes from 35,3% to 28,8% of the Brazilian population. In 2003, poverty still reached 28,2% of the population having grown during Lula’s first year in government, as we announced first in 2004. In 2003, a new period of fall begins, reaching 22.7%. This amounts to a cumulative decrease of 19,18% between 2003 and 2005, comparable to the decrease of. 39.

(40) 18,47% between 1993 and 1995. The existing parallel in the poverty reduction between the two episodes ten years apart may be seen in the graph below.. Proportion of Poor in the Population (% in class E). Source: CPS/FGV based on PNAD microdata/IBGE. In 2006, inspired by this leap changes in the pattern of poverty, we launched a research with a provocative title “Second Real”. Since then, poverty has kept its descending trend, falling almost 30% since 2005. Following the metrics imposed by the immediate effects of the Real Plan on poverty that we had the pleasure to detect first 13 years ago, today there would be three reais of poverty reduction, for the 2003 to 2008 period when poverty fell 43%. Poverty reduction in the southeastern and northeastern regions is in the graph below:. 40.

(41) Poverty Rates (% Class E) Northeastern and southeastern regions. Source: CPS/FGV based on PNAD microdata/IBGE. Below we present the map of different states between these two periods of marked poverty reduction between 1993 to 1995 and 2003 to 2008, expressed in annual rates of poverty reduction to enable a comparison between them. In Brazil, between 1993 and 1995, poverty decreased 9,6% per year, while the annual average of the last five years was 10,6%. In regional terms, except for some states such as Amazonas, Acre, Roraima and Rio de Janeiro, the annual pace of poverty reduction now tends to be stronger in other states. In the case of Rio de Janeiro State, the fact that the Real boom through exchange rate valuation may have benefited the non-transaction sector may explain the phenomenon, given the relative importance of the service sector in the state (vide Neri 1996). 41.

(42) Annual Rate of Poverty Reduction (Class E) – 2003/2008 8 0 0 2 a 3 0 0 2 ] E e s s a l c [ a z e r b o p a d o ã ç u d e r e d l a u n a a x a T l a u n A a x a T. % 5 é t A. % 5 , 7 a % 5 e d. % 0 1 a % 5 , 7 e d. % 5 , 2 1 a % 0 1 e d. % 5 1 a % 5 , 2 1 e d. % 5 1 e d s i a M. Annual Rate of Poverty Reduction (Class E) – 1993/1995. l a u n A a x a T. % 5 é t A. % 5 , 7 a % 5 e d. % 0 1 a % 5 , 7 e d. % 5 , 2 1 a % 0 1 e d. % 5 1 a % 5 , 2 1 e d. % 5 1 e d s i a M. 42.

(43) Local Contribution to Poverty Reduction (Size of class E) In order to complement the analysis of the relative annual changes, we present the contribution of each local area to the rise in classes ABC in the recent period. Next, we present the contribution of each geographic unit to the reduction of poverty in 2007 to 2009 and 2003 to 2008. In both periods, the Northeast stands out with 44,28% and 44,70% of poverty reduction observed during the respective time periods. From 2003 to 2008, eight million people crossed the poverty line in the northeast.. Contribution of Spatial Units to Poverty Reduction 2007-08 and 2003-08 In numbers of people and proportion of the total poverty decrease Contribution Population Population Category. 2007-2008. Total. 3800837. Contribution % 2003-2008 19454189. Population Category. 2007-2008. North Northeast Southeast South Center. 330147 1683090 1445943 184478 151100. 2007-2008. Capital Periphery of metropolises (non capital) Urban area – non metropolitan Rural area. 1055055 338641 1582543 823512. 1441725 8696888 6233898 1606360 1476818. 2007-2008. Rondônia Acre Amazonas Roraima Pará Amapá Tocantins Maranhão Piauí Ceará Rio Grande do Norte Paraíba Pernambuco Alagoas Sergipe Bahia Minas Gerais. 20751 13366 84760 3347 76648 51267 80932 263454 134136 389980 127081 139379 170441 -39036 34211 466707 333060. 43. 100.00%. 2007-2008. 2003-2008. 8.69% 44.28% 38.04% 4.85% 3.98%. 7.41% 44.70% 32.04% 8.26% 7.59%. Contribution 2003-2008 4504513 2041645 8547525 4372630. Population Category. 2003-2008. 100.00% Contribution %. 2003-2008. Population Category. 2007-2008. 2007-2008. 2003-2008. 27.76% 8.91% 41.64% 21.67%. 23.15% 10.49% 43.94% 22.48%. Contribution % 2003-2008 91142 46853 299776 28700 636971 108341 230188 1178375 531964 1324724 512316 598443 1346262 492719 249570 2466365 2175137. 2007-2008. 2003-2008. 0.55% 0.35% 2.23% 0.09% 2.02% 1.35% 2.13% 6.93% 3.53% 10.26% 3.34% 3.67% 4.48% -1.03% 0.90% 12.28% 8.76%. 0.47% 0.24% 1.54% 0.15% 3.27% 0.56% 1.18% 6.06% 2.73% 6.81% 2.63% 3.08% 6.92% 2.53% 1.28% 12.68% 11.18%.

(44) Espírito Santo Rio de Janeiro São Paulo Paraná Santa Catarina Rio Grande do Sul Mato Grosso do Sul Mato Grosso Goiás Distrito Federal. 48847 542839 523260 55320 -55370 184219 2571 58607 83564 7088. 456818 876573 2723409 782092 195439 630806 214971 386690 653518 222025. Population Category. 2007-2008. RO Capital AC Capital AM Capital RR Capital PA Capital PA Periphery AP Capital TO Capital MA Capital PI Capital CE Capital CE Periphery RN Capital PB Capital PE Capital PE Periphery AL Capital SE Capital BA Capital BA Periphery MG Capital MG Periphery ES Capital RJ Capital RJ Periphery SP Capital SP Periphery PR Capital PR Periphery SC Capital RS Capital RS Periphery MS Capital MT Capital GO Capital DF Capital. 15334 6288 142513 -1071 9089 22591 36105 14103 19158 27579 30969 29453 41617 -1497 24086 1576 -40099 14146 174468 -28139 16850 1297 -6206 383180 72638 106645 277450 -14834 -39454 -3164 36887 1687 -1159 -1320 18948 7088. 2.35% 4.51% 14.00% 4.02% 1.00% 3.24% 1.11% 1.99% 3.36% 1.14%. Contribution 2003-2008 33981 24646 268852 19494 215985 82464 60286 35241 230825 151733 305433 186656 128217 101129 193137 287817 146914 66478 423684 139357 159592 195800 20455 326372 304940 970601 607870 102116 115266 15901 65207 121773 69715 41990 103044 222025. Source: CPS/FGV based on PNAD microdata/IBGE. 44. 1.29% 14.28% 13.77% 1.46% -1.46% 4.85% 0.07% 1.54% 2.20% 0.19%. 2007-2008. 2003-2008. 0.40% 0.17% 3.75% -0.03% 0.24% 0.59% 0.95% 0.37% 0.50% 0.73% 0.81% 0.77% 1.09% -0.04% 0.63% 0.04% -1.05% 0.37% 4.59% -0.74% 0.44% 0.03% -0.16% 10.08% 1.91% 2.81% 7.30% -0.39% -1.04% -0.08% 0.97% 0.04% -0.03% -0.03% 0.50% 0.19%. 0.17% 0.13% 1.38% 0.10% 1.11% 0.42% 0.31% 0.18% 1.19% 0.78% 1.57% 0.96% 0.66% 0.52% 0.99% 1.48% 0.76% 0.34% 2.18% 0.72% 0.82% 1.01% 0.11% 1.68% 1.57% 4.99% 3.12% 0.52% 0.59% 0.08% 0.34% 0.63% 0.36% 0.22% 0.53% 1.14%.

(45) Economic classes Incidentally, our classification of economic classes (classes E, D, C and AB) derives from our definition of poverty conceived as class E. So the basis of the income pyramid was already addressed in the topic above. An interesting synthesis of the other extreme of the distribution may be found in the sum of classes ABC. The next table presents the absolute and relative sizes of the population in various economic classes. It is interesting to note that within PNAD, class C is not only prevalent in terms of population size encompassing almost of the Brazilian population (49,22% or 91 million people) in 2008, and that if the increase of 2,5% in the post-crisis period seen on PME is confirmed by next year’s PNAD, not only the average voter, that is the one who decides the election would now already be a part of class C – but also class C alone would decide the election in the first round. Besides possibly crossing the threshold of 50%, class C according to our estimates, would become the dominant class in terms of income mass in 2008, with 45,75% of income versus 44,05% of class AB. Therefore, together classes ABC would hold 89,8% of the income accounted for in PNAD.. 2008. TOTAL Class E Class D Class C Class AB. Population Income Mass Total (people) % Pop Reais % Mass 186.440.290 100,00% 110.395.816.985 100,00% 29.860.927 16,02% 2.228.819.591 2,02% 45.399.117 24,35% 9.030.338.362 8,18% 91.762.175 49,22% 50.506.818.742 45,75% 19.418.071 10,42% 48.629.840.290 44,05% Source: CPS/FGV based on PNAD microdata/IBGE. Local classes The leader in average income (R$ 1249 by month) and in participation of classes ABC is the city of Florianópolis (92,6% of its population). Curitiba with 86,49% is the second capital with the largest number of people in class ABC (86,5%), followed by Vitoria (80%). Rio has 75,3% in classes ABC (7th among the 27 capitals) and each carioca earns 1015 reais (6th largest). In the opposite extreme, we find the peripheries of Fortaleza (40,08%), Salvador (42,42%) and Recife (42,89%).. 45.

(46) rank 2008. % Classe ABC %. %. 2008. 2007. rank. % rank. 2003. Var (%) rank. Var (%). 2007/2008 rank. 2003/2008. 1 2 3 4 5. Florianópolis - SC Curitiba - PR Vitória - ES Porto Alegre - RS Belo Horizonte - MG. 92.61 86.49 80.07 80.05 78.61. 1 2 3 6 5. 90.38 84.97 82.18 75.29 75.60. 1 4 2 3 7. 81.82 71.43 73.57 72.08 62.00. 26 29 36 17 22. 2.47% 1.79% -2.57% 6.32% 3.98%. 34 28 36 35 21. 13.19% 21.08% 8.84% 11.06% 26.79%. 1 2 3 4 5. Periferia de Fortaleza - CE Periferia de Salvador - BA Periferia de Recife - PE Periferia de Belém - PA Maceió - AL. 40.08 42.42 42.89 45.28 47.61. 36 34 35 33 32. 34.56 42.72 39.29 42.99 44.62. 36 35 34 33 32. 19.45 24.12 28.91 29.87 32.45. 3 34 12 19 16. 15.97% -0.70% 9.16% 5.33% 6.70%. 1 2 9 7 11. 106.07% 75.87% 48.36% 51.59% 46.72%. Source: CPS/FGV based on PNAD microdata (acesse anexo 1.1 ou http://www.fgv.br/ibrecps/RET4/CPC_evolucao_espacial/). When we assess the total figures of capitals and peripheries together with the remaining sizes of cities, as expected, we found the lowest number of people in class ABC in rural areas (34,96% versus 68,48% in capitals). In terms of advancements in last year, both capitals and metropolitan peripheries present growth slightly superior to the country’s level. Just urban areas outside metropolises presented lower growth than the national level. % Classe ABC. Total Capital Metropolis periphery (non capital) Urban area - non-metropolitan Rural area. %. %. %. 2008 59.64 68.48 63.97 61.07 34.96. 2007 56.63 64.82 60.66 58.47 32.83. 2003 45.16 54.62 49.47 46.31 21.98. Var (%) 2007/2008 5.32% 5.65% 5.46% 4.45% 6.49%. Var (%) 2003/2008 32.06% 25.38% 29.31% 31.87% 59.05%. Source: CPS/FGV based on PNAD microdata (acesse anexo 1.1 ou http://www.fgv.br/ibrecps/RET4/CPC_evolucao_espacial/). Just like its capital city Florianopolis, Santa Catarina State also leads the ABC ranking with 82,32% of population in this income group. The remaining southern states complete the first positions in the rankings: Paraná with 79,85% and Rio Grande do Sul with 73,29%. Meanwhile, in the opposite extreme, we found only northeastern states, Maranhão (32,22%), Alagoas (32,25%) and Piauí (36,99%).. 46.