The New Middle-Class
Coordination:
Marcelo Cortes Neri
August 2008
1
The authors take full responsibility for the ideas published here. The opinions expressed do not
necessarily reflect Fundação Getulio Vargas´ point of view.
The New Middle Class (A Nova Classe Média) / Marcelo Côrtes Neri. Executive
Summary / Marcelo Côrtes Neri (coord). - Rio de Janeiro: FGV/IBRE, CPS, 2008.
[70] p.
1. Middle-class 2. Income 3. Poverty 4. Inequality 5. Social mobility. I.
Neri, M.C. II. Fundação Getulio Vargas, Instituto Brasileiro de Economia. Centro
de Políticas Sociais.
© Marcelo Neri 2008
2
The New Middle-Class
Rio de Janeiro, August 05, 2008
Version 2.0: August 8, 2008
Centro de Políticas Sociais
Instituto Brasileiro de Economia
Fundação Getulio Vargas
Coordination:
Marcelo Cortes Neri
[email protected]
CPS’ Staff:
Luisa Carvalhaes Coutinho de Melo
Samanta dos Reis Sacramento
André Luiz Neri
Carolina Marques Bastos
Celio Maymone Pontes
Ana Lucia Salomão Calçada
Ana Beatriz Urbano Andari
Celso Henrique Fonseca
3
Table of Contents
1. Motives
2. Methods
3. Mean income, inequality and poverty
4. Defining the middle-class
5. Changes in the distributive pie
6. The distributive dance
7. The return of the C from “carteira de trabaho” (employment
registration booklet)
8. Conclusions
9. Bibliography
Appendix I – Portrait of the Middle-Class
Appendix II – Regional Variation
Appendix III – Multivariate Exercises
4
The New Middle-Class1
August 2008
1. MOTIVES
Brazil received an “investment grade” by the international rating agencies in the first
semester of 2008. In 2007, Brazil also joined a group of countries with a high Human
Development Index (HDI). The present study reveals these events’ daily effects on the
human anatomy’s most sensitive part: the pocket. Class “C” share increased 22.8%
from April 2004 to April 2008, while classes “A” and “B” grew 33.6% in the same
period. Therefore, if one assumes that the middle-class is richer than Brazil’s class “C”,
the conclusion that the middle-class has grown remains unaltered. Another point is that
all substantive and symbolic indicators show the occurrence of a boom in class “C”:
access to housing, car, computer, credit and formal employment are at their historical
highs.
Class “C” is the central class, below A and B e above D and E. In order to quantify
these levels, we calculated the household per capita income from work and then express
it in terms equivalent to the total household incomes from all sources. The central Class
C refers to household incomes between R$ 1064 and R$ 4561 at today’s prices from
São Paulo Metropolitan region. Class C is immediately above the 50% poorest and the
10% richest at the turn of the century. Illustratively, class C is limited by the Indian
border on one side and to the Belgian border on the other side in the so-called Belíndia.
Some see class C as the low-middle class, and class B as the upper-middle class. We are
not the USA, but this is also a free country. It is more important to have a well-defined
and consistent criterion. Our class C receives the average income in society, that is, it is
the middle class in the statistical sense. Given the income inequality, the average
1
Two years ago, Fundação Getulio Vargas president, Mr. Carlos Ivan Simonsen Leal, proposed that we
went beyond the statistics relating to poverty and studied the Brazilian middle-class. This study is a direct
consequence of this challenge; hence we would like to thank his guidance and encouragement. The
responsibility for the methodological choices and the results of the research rests with the author.
5
income is high in relation to our median. In relation to the rest of the world: 80% of the
people in the world live in countries with lower per capita income levels than the
Brazilians. Our Class C is thus the closest image to the Brazilian society.
If we want to understand Brazil, it is necessary to quit looking only at the rear-view
mirror that focuses on obsolete data, and start looking ahead through the windshield of
the current Brazilian scene. Even before looking ahead and forecasting, given the
uncertainty as a result of the current American crisis, it is useful to explore more recent
data. Here, we inform the public debate by providing unpublished data on 2007 and
2008. As a function of the continuous improvement in the social conditions with the
successive reductions in inequality and poverty rates presented here, the study unveils
the emergence of a new middle-class in Brazil. This research defines and details the
economic rise of this new middle-class in the main Brazilian cities. The emphasis will
be on income from work in the six main Brazilian cities as a result of the greatest data
availability of recent Monthly Employment Survey (PME/IBGE) microdata (up to April
2008). Apart from being more recent, this data also allow us to monitor the same
families through time and to observe their transitions into- and out of poverty and
middle-class statuses. Roughly, while the traditional studies based on PNAD, and even
on the PME, provide static snapshots of society groups in different moments, the PME
also makes it possible to capture the trajectory of the same families, isolating the
emerging ones from those that already belonged to the middle-class, or alternatively,
identifying the new destinations of those who were living in old-time poverty. In order
to make the current scenario as complete as possible, we also count on data from
Caged/MTE (Cadastro Geral de Emprego e Desemprego do Ministério do Trabalho e
Emprego – General Record of Employment and Unemployment from the Ministry of
Labour and Employment) available for all national territory until June 2008. This data
refers to formal employment rates, which have reached a record high in the last months.
The so-called return of the employment registration booklet (carteira de trabalho) may
be the most symbolic element in the reappearance of the new Brazilian middle-class.
Overall, much has been said in this decade in terms of inequality (since 2001) and
poverty reduction (since 2004); the role of official income transfer programs to the
poorest has been emphasized; but little attention has been paid to the structural advances
6
that resulted from the labor expansion observed for all layers of society. Since 2004
until now, there has been an increase in the income from work in general and in the
creation of formal job posts, in particular. And since the last statistical snapshot on
Brazil, taken through the PNAD 2006 lenses, income from work has been its most
striking feature. This research, based on most recent data, shows the continuity of this
phenomenon coupled with some acceleration in the pace of meaningful decreases in
poverty and inequality in Brazil until now. After two lost decades in terms of income
and labor, the combination of faster growth with a marked reduction in inequality for a
longer period is remarkable – making this a story worth telling, especially for being
unprecedented in the documented history of Brazilian statistics.
After years of crises in metropolitan areas, the main finding concerns the growth of a
middle-class and the reduction of poverty in these same areas; and how much of it can
be attributed to the growth in population’s income from private work. In many cases,
the analyses give emphasis either to the role of large public transfers of income to the
population, as an expansion of the Bolsa Família, or the social security transfers,
contributory and non-contributory, related to adjustments in the minimum wage. We
argue that, since 2004 at least, an increase in income from work rivals with those
transfers in explaining the improvements in income for the population as a whole (and
that since 2001, for the poorest segments in the population). This is well documented in
the PNAD series until October 2006. The recent labor return and the absolute growth of
the middle groups of the population have not yet been considered in detail.
Beside the prevalence of private earnings derived from work over public transfers; the
metropolitan areas vis-à-vis the rest of the country; and the emergence of the middleclass; the change in the international context must also be understood. Until mid-2007,
despite a stronger growth in the Brazilian GDP since 2004 compared to the previous
two lost decades, Brazil was chasing the growth rates of central and emerging
economies such as India and China. Since then, Brazilian GDP caught up with these
rates in the international context. As it has been reported by the main international
newspapers, Brazil is in evidence in terms of growth. This paper substantiates this new
sense of prosperity through objective data on the improvement in the aspects that were
central to the previous crises, namely: income from work, formal employment,
7
metropolitan areas and the so-called middle-class. Another novelty well spotted by Mac
Margolis from Newsweek in his blog is that for the first time in more than twenty years
living in Brazil, the word crisis is pronounced here referring to somewhere else.
The present paper explores the movements in income distribution in the statistical sense,
encompassing changes in inequality, as well as in the average growth of household
income. We give special emphasis to the analysis of two phenomena, namely: poverty
at the lower end of per capita income distribution; and, the so-called middle-class at the
center of total household income. The work plan is as follows: in the second section, we
analyze the recent evolution of the distribution indicators based on per capita income
from work; in the third section, we detail the regional evolution of indicators such as the
median, average and poverty rates; in the fourth chapter, we take advantage of PME’s
longitudinal aspect in order to assess the frequency of movements into- and out of
poverty, which are associated with the minimum wage movements in the following
section. The sixth and last sections present the main conclusions for this study.
8
9
2. METHOD
It is traditional amongst the research institutions like IBGE to use data from the
Monthly Employment Survey (PME) at the individual’s level, as opposed to the
households’ level, particularly placing focus on indicators such as unemployment rate
and the average income from work. Nevertheless, PME is a household survey
comparable to the National Household Survey (PNAD) and can be used as such. This is
noteworthy because the evaluation of the socio-economic conditions must take into
consideration the process of allocating resources within the household. For instance,
how income from an adult worker may benefit other members of the family, such as the
children. In this sense, the most adequate variable to measure the level of poverty would
be the household per capita income, which is the sum of income from all members of
the household divided by the total number of members. Similarly, when trying to define
the size of the middle-class in order to assess the purchasing power of family goods
such as a house, the adequate variable is the total income received by all household
members. Both concepts sum up a series of factors that influence all members of a
family, such as the level of occupation and income, measured formally or informally,
but whose effects can be shared by the total number of household members (BARROS
et al.,1996).
The main question here is how to improve the monitoring of our population’s living
conditions. How do we assess the social and economic performance based only on the
data from PNAD, whose information is on average 18 months old? For instance, it has
been almost two years since the last national snapshot based on PNAD. The increase in
speed is a necessary requirement to design an effective social target evaluation system.
This includes managerial systems within public administration, as well as the
monitoring of poverty levels by society itself. This is no less urgent from the standpoint
view of private companies that want to adapt their business cycle fluctuations in order
to target their demand. In view of these needs, we propose to process the PME
microdata, which – given its agile nature – allows us to reduce that 18-month lag down
to 3 months (NERI; CONSIDERA, 1996).
Apart from this benefit derived from an increased speed in the dissemination of
information, using PME/IBGE data on a monthly basis allows capturing the determinant
factors of the distribution of income from work, observed in Brazil through time. The
10
series of average household per capita income and of inequality captured by the Gini
index, presented on graphs 1 and 2 below, and detailed ahead, indicate that the greatest
part of the growth of per capita income from work among the poorest classes observed
in the last four years occurred between March and June 2004; but it remains
uninterrupted ever since.
GRAPH 1 - Poverty Series
38
37
36
35
34
33
32
31
30
29
28
27
26
25
24
Mar-08
Dec-07
Sep-07
Jun-07
Mar-07
Dec-06
Sep-06
Jun-06
Mar-06
Dec-05
Sep-05
Jun-05
Mar-05
Dec-04
Sep-04
Jun-04
Mar-04
Dec-03
Sep-03
Jun-03
Mar-03
Dec-02
Sep-02
Jun-02
Mar-02
Source: CPS/IBRE/FGV based on PME/IBGE microdata.
Criterion: Usual per capita income
GRAPH 2 - Gini’s Index Inequality Evolution
0,64
0,63
0,62
0,61
0,60
0,59
0,58
Mar-08
Dec-07
Sep-07
Jun-07
Mar-07
Dec-06
Sep-06
Jun-06
Mar-06
Dec-05
Sep-05
Jun-05
Mar-05
Dec-04
Sep-04
Jun-04
Mar-04
Dec-03
Sep-03
Jun-03
Mar-03
Dec-02
Sep-02
Jun-02
Mar-02
Source: CPS/IBRE/FGV based on PME/IBGE microdata.
Criterion: Usual per capita income
Roughly, if social scientists were astronomers and the income distribution were a
celestial body in movement, PNAD would be like a super-telescope in the right place to
capture that body’s passage. Nonetheless, we need more than snapshots of the
11
phenomenon in several years, such as those given by PNAD, in order to precise the
factors that determine the trajectory of social indicators based on income, such as
poverty and inequality. It would be necessary to use a sort of film based on a series of
monthly snapshots, such as those from the PME, which allows us to identify the effect
of discreet changes on policy variables, such as abrupt changes in interest rates, in the
exchange rate or, more bluntly, the role of minimum wage adjustments (NERI, 1995).
Thirdly, PME uses a rotating panel methodology similar to that adopted by the Current
Population Survey (CPS) in the US, which monitors the information of the same
individuals and their families during consecutive observations. In other words, in our
analogy with the movies, we are not only following the story of sub-groups in society,
but developing a film for each person in the sample. In particular, we explore data
collected in four consecutive months (example: March, April, May and June in 2008
given year) and so forth – as this period is especially interesting to identify the effects of
the minimum wage in each year and the rapid reduction in poverty and inequality
occurred in 2004, as mentioned above. The approach used in this work consists in
calculating the likelihood of transition into - and out of poverty (or of middle class), as
well as the non-transition within four consecutive months. The individuals most
affected will be identified through the microdata, which will also help to reconstruct
their respective household per capita work income trajectories (BARROS et al., 1996).
Finally, it is important to highlight two PME limitations, as follows: it does not consider
income unrelated to work, such as those from government income transfer programs
and income from interest gains for the groups with a financial wealth stock; it only
covers the six main metropolitan areas in Brazil. In short, this research only provides
evidence of labor in the metropolitan areas. (RAMOS; BRITO, 2003). Our
methodology will address this limitation.
12
3. MEAN, INEQUALITY AND POVERTY
MEAN
The usual or normal income concept smoothes transitory fluctuations of income, such as
the income from the so-called “décimo terceiro salário” (an extra monthly wage paid at
Christmas time), holiday bonus and overtime work. The concept of effective income
also surveyed by the PME presents marked seasonal fluctuations as the year begins as
shown by the graphs, but – these peaks aside – the data series are relatively close2. We
have chosen to work with the usual concept of income not only because it eliminates
erratic fluctuations – which can create an upward bias towards the mobility measures –
but it is also advantageous because the PNAD uses it, enabling a direct comparison
between the results with the main database of the Brazilian household survey system.
GRAPH 3 - Evolution of Per capita average income
750
725
700
675
650
625
600
575
550
525
500
475
Usual
Mar-08
Dec-07
Sep-07
Jun-07
Mar-07
Dec-06
Sep-06
Jun-06
Mar-06
Dec-05
Sep-05
Jun-05
Mar-05
Dec-04
Sep-04
Jun-04
Mar-04
Dec-03
Jun-03
Sep-03
Mar-03
Dec-02
Jun-02
Sep-02
Mar-02
450
Effective
Source: CPS/IBRE/FGV based on PME/IBGE microdata.
Criterion: Usual per capita income
A first approach is to look at the evolution of the income distribution in the statistical
sense, including the evolution of the usual per capita income average and inequality.
We have decided to present the graphs with a moving average in order to better isolate
the trends. The income average gives continuity to the expanding trajectory observed at
the end of the 2003 recession, as the monthly data have already suggested. The average
per capita income growth rate between April 2003 and April 2008 – thus discounting
2
Neri (1996) details the differences among the two concepts based on a comparison between PME
collected in 1980 and 1982, before its reformulation. Another point is that the usual concept tends to be
linked to the month of the survey, while the effective concept is linked to the previous one. In this sense,
the effective concept would be more appropriate as a leader indicator in the PNAD. On the other nad, the
effective concept of income is the most adequate to the PME series between 1982 and 2002, before its
second reformulation
13
the population growth – is 5% per year. This pace is not inferior, for instance, to that
observed during the economic miracle in the 1960s’. Once the post-April 2004 period is
isolated, this rate reaches 6,5% per year, once again having discounted the population
growth.
GRAPH 4 - Evolution of Per capita average income
12-month moving average
600
575
550
525
500
475
Source: CPS/IBRE/FGV based on PME/IBGE microdata.
Criterion: Usual per capita income
TABLE 1 - Per Capita Income – 15 to 60 years old
(6 Brazilian metropolitan areas)
Level (R$)
PER CAPITA INCOME– R$
abr/02
abr/03
abr/04
abr/05
abr/06
abr/07
abr/08
514,85
480,51
467,47
513,04
536,07
574,69
605,42
Evolution
PER CAPITA INCOME
Moving
Average
Variation(%) Variation(%) Difference(R$)
Annual Evolution (12 meses)
abril 03 / abril 02
abril 04 / abril 03
abril 05 / abril 04
abril 06 / abril 05
abril 07 / abril 06
abril 08 / abril 07
-6,67
-2,71
9,75
4,49
7,20
5,35
-9,20
8,78
3,87
7,39
4,14
-34,34
-13,04
45,57
23,03
38,62
30,73
14,75
26,37
16,17
11,84
4,14
124,91
137,95
92,38
69,35
30,73
Accumulated Evolution (desde 2002)
abril 08 / abril 03
abril 08 / abril 04
abril 08 / abril 05
abril 08 / abril 06
abril 08 / abril 07
26,00
29,51
18,01
12,94
5,35
Criterion: Per capita Usual Income
Source: CPS/IBRE/FGV based on PME/IBGE microdata.
14
Feb-08
Nov-07
Aug-07
May-07
Feb-07
Nov-06
Aug-06
May-06
Feb-06
Nov-05
Aug-05
May-05
Feb-05
Nov-04
Aug-04
Feb-04
May-04
Nov-03
Aug-03
Feb-03
May-03
450
Social classes, in general, are typically defined based on the total household income and
not per capita income, as in the case of poverty studies. In relation to the household total
income, there was a decline in the two first years, especially between April 2002 and
April 2003, with a decrease of 9% in the total real income. The highest level in the
whole series (R$ 1957 in 2008) occurred in 2008, 24,76% above the 2004 level in real
terms.
TABLE 2 - Household Income – 15 to 60 years old
(6 Brazilian metropolitan areas)
Level (R$)
HOUSEHOLD TOTAL INCOME - R$
abr/02
abr/03
abr/04
abr/05
abr/06
abr/07
abr/08
1784,08
1628,11
1568,47
1704,74
1770,08
1886,36
1956,90
Evolution
HOUSEHOLD TOTAL INCOME
Moving
Average
Variation (%) Variation (%) Difference (R$)
Annual Evolution(12 months)
abril 03 / abril 02
abril 04 / abril 03
abril 05 / abril 04
abril 06 / abril 05
abril 07 / abril 06
abril 08 / abril 07
-8,74
-3,66
8,69
3,83
6,57
3,74
-10,00
7,91
3,31
6,87
2,82
-155,97
-59,64
136,27
65,34
116,28
70,54
10,25
22,50
13,53
9,89
2,82
172,82
328,79
388,43
252,16
186,82
70,54
Accumulated Evoluation since 2002
abril 08 / abril 02
abril 08 / abril 03
abril 08 / abril 04
abril 08 / abril 05
abril 08 / abril 06
abril 08 / abril 07
9,69
20,19
24,76
14,79
10,55
3,74
Criterion: usual income
Source: CPS/IBRE/FGV based on PME/IBGE microdata.
15
INEQUALITY
Before the analysis of inequality measures, it is worth emphasizing that, apart from the
geographical coverage and the income concept used, the per capita income concept used
includes the null income, which is fundamentally important to the inequality measures.
The most used measurement is the Gini index that varies from 0 to 1 and assigns weight
to income in an inverse order of their ranking. That is, the richest person in a society is
given the smallest weight, which increases as we move towards the smallest income
levels. In this sense, the person with a 0 income should be awarded the biggest weight
and not the smallest, null weight - when we would instead implicitly disregard them in
the analysis. These methodological restrictions notwithstanding, income inequality also
presents a noticeable retraction throughout the series. The Gini index falls from 0,627 in
April 2002 to 0,584 in April 2008, which is considerable given the scale of variation in
the index, particularly within the Brazilian context. The Gini index of per capita income
from all sources remains stagnated around 0,6 between the 1970 and 2000 Census. The
only impressive permanent change observed in the statistically documented series in the
country was the famous increase in the 1960’s when the Gini index for individual
income increased 0,07 point in a decade. Conceptual and geographical differences aside,
for comparison purposes, this absolute decrease in six years 0,0426 is exactly in the
same rhythm in the 1960s’.
GRAPH 5 - Inequality Evolution – Gini Index
12-month moving average
0,635
0,630
0,625
0,620
0,615
0,610
0,605
0,600
0,595
0,590
Feb-08
16
Nov-07
Criterion: Usual per capita income
Aug-07
May-07
Feb-07
Nov-06
Aug-06
May-06
Feb-06
Nov-05
Aug-05
May-05
Feb-05
Nov-04
Aug-04
May-04
Feb-04
Nov-03
Aug-03
May-03
Feb-03
Source: CPS/IBRE/FGV based on PME/IBGE microdata.
TABLE 3 - Labour Inequality – Total Population
(6 Brazilian metropolitan areas)
Level
GINI
Apr/02
Apr/03
Apr/04
Apr/05
Apr/06
Apr/07
Apr/08
0,6270
0,6284
0,6258
0,6036
0,6011
0,5963
0,5844
Evolution
GINI
M. Móvel
Variation (%) Variation (%) Difference (p.p.)
Annual Evolution(12 meses)
April 03 / April 02
April 04 / April 03
April 05 / April 04
April 06 / April 05
April 07 / April 06
April 08 / April 07
0,22
-0,40
-3,56
-0,41
-0,80
-2,00
-2,17
-2,10
-1,12
-0,82
-1,33
0,001
-0,003
-0,022
-0,002
-0,005
-0,012
-7,32
-5,26
-3,23
-2,13
-1,33
-0,043
-0,044
-0,041
-0,019
-0,017
-0,012
Accumulated Evolution (since 2002)
April 08 / April 02
April 08 / April 03
April 08 / April 04
April 08 / April 05
April 08 / April 06
April 08 / April 07
-6,79
-7,00
-6,62
-3,18
-2,78
-2,00
Criterion: per capita usual income
Source: CPS/IBRE/FGV based on PME/IBGE microdata.
POVERTY
Income distribution and labor poverty improve the function of the average income
accelerated growth , as well as the reduction in income inequality. Here, we present
estimates of metropolitan areas labor poverty using the poverty line of 135 Reais per
month per person at Sao Paulo metropolitan area prices, adjusted for spatial differences
in living costs, according to the Center for Social Policies at the Fundação Getulio
Vargas (Neri 2006 e Ferreira, Lanjouw e Neri 2003).
17
GRAPH 6 - Labor Poverty Series
12-month moving average
37
36
35
34
33
32
31
30
29
28
27
Source: CPS/IBRE/FGV based on PME/IBGE microdata.
Criterion: Usual per capita income
TABLE 4 - Poverty – Total Population
(6 Brazilian Metropolitan areas)
Level
Class E
Apr/02
Apr/03
Apr/04
Apr/05
Apr/06
Apr/07
Apr/08
34,93
37,13
37,17
32,58
31,61
29,09
25,16
Evolution
Class E
Moving
average
Variation (%) Variation (%)
April 03 / April 02
April 04 / April 03
April 05 / April 04
April 06 / April 05
April 07 / April 06
April 08 / April 07
6,31
0,11
-12,34
-2,99
-7,97
-13,50
Accumulated
Evolution
April 08 / April 02
April 08 / April 03
April 08 / April 04
April 08 / April 05
April 08 / April 06
3,60
-10,74
-5,81
-6,61
-7,42
Difference
(p.p.)
2,20
0,04
-4,59
-0,98
-2,52
-3,93
(since 2002)
-27,97
-32,24
-32,31
-22,78
-20,39
-24,70
-27,31
-18,56
-13,54
-9,77
-11,97
-12,01
-7,42
-6,45
Criterion: Per capita usual income
Source: CPS/IBRE/FGV based on PME/IBGE microdata.
18
Feb-08
Nov-07
Aug-07
May-07
Feb-07
Nov-06
Aug-06
May-06
Feb-06
Nov-05
Aug-05
May-05
Feb-05
Nov-04
Aug-04
May-04
Feb-04
Nov-03
Aug-03
May-03
Feb-03
26
Labor Poverty – 15 to 60 years
TABLE 5 - Miséria Trabalhista – 15 a 60 anos
6 Regiões Metropolitanas brasileiras
Nível
Class E
abr/02
abr/03
abr/04
abr/05
abr/06
abr/07
abr/08
28,64
30,46
30,45
25,42
24,55
21,72
18,39
Evolução
Class E
Variation (%) Variation (%) Variation (%)
Annual Evolution (12 meses)
abril 03 / abril 02
abril 04 / abril 03
abril 05 / abril 04
abril 06 / abril 05
abril 07 / abril 06
abril 08 / abril 07
6,34
-0,04
-16,51
-3,42
-11,54
-15,30
3,71
-14,09
-7,35
-8,96
-9,78
1,82
-0,01
-5,03
-0,87
-2,83
-3,32
-32,20
-34,62
-23,90
-17,87
-9,78
-10,25
-12,06
-12,05
-7,03
-6,16
-3,32
Accumulated Evolution (since 2002)
abril 08 / abril 02
abril 08 / abril 03
abril 08 / abril 04
abril 08 / abril 05
abril 08 / abril 06
abril 08 / abril 07
-35,78
-39,61
-39,59
-27,64
-25,08
-15,30
Critério: Renda Per Capita Habitual
Source: CPS/IBRE/FGV based on PME/IBGE microdata.
19
4. DEFINING THE MIDDLE-CLASS
Defining the middle-class is like defining an elephant, if you have never seen one, it is
hard to visualize it. There are at least two perspectives to conceptualize the class. One is
through the analysis of people’s attitudes and expectations. Brazilian consumers’
survey, as released monthly by the Brazilian Institute of Economics at the Fundação
Getúlio Vargas, goes in that direction. This kind of approach was much developed in
the 1950s’ and 1960s’ by George Katona, a behavior psychologist who greatly admired
James Tobin. Following the same line, Thomas Friedman, international columnist at the
New York Times in his best-seller book “The world is flat”, defines middle-class more
than merely by its present living conditions and income, but also for its hoping to be in
a better position in the future. This upward social structural mobility would be just like
realizing the American dream, understood as the possibility of social ascension in each
country. The class is that group that shares a plan of individual progress.
Additionally, we propose (but do not release here) the use of measures of life quality
and other subjective questions from the new line of surveys as the Gallup World Poll,
IPSOS similar, whose advantage is its high international comparability for applying the
same questionnaire to a great number of countries. This advantage is also shared by the
regionally based surveys, such the LatinBarometer in Latin America and the
EuroBarometer in the old continent. In particular, we suggest the use of direct measures,
such as the expectation of happiness five years into the future in comparison with the
current level of happiness – by asking questions where the person may attribute a
subjective grade from 0 to 10 about their respective satisfaction with life. This kind of
analysis was used in the Future Happiness Index (FHI) that we developed for a project
with the Inter-American Development Bank, based on a sample of more than 132
countries covered by the 2006 Gallup World Poll. This index will be released shortly.
For now, we can say that the data indicate, measured by the difference between the
current and future happiness levels, is high in relation to other countries which is
consistent with the enlarging Brazilian middle class discussed here.
The second way to define the social classes (E, D, C, B2, B1, A2 and A1) is according
to their consumption potential, such as the Brazil Criterion in which the middle-class (r
for some the lower middle class) is called Class C. This stratification is implemented
according to the impact of access to durable goods and their respective amount (TV set,
20
radio, washing machine, fridge, freezer, DVD or video player), toilettes, servants and
level of education of the head of the household3. This criterion estimates the weights
from a mincerian equation (log of the total household income). CPS proposes a
complementary conceptualization to measure the evolution of the new middle class in
Brazil, also from a producer’s point of view – that is, the capacity to keep this
consumption potential through time. In this yet unpublished work, apart from testing the
measurement of the middle class based on a combination of elements such as income
and access to consumer goods traditionally used, we also propose to measure the middle
class based on the capacity of income generation and wealth sustainability in the longer
term. In the first aspect, there is access to public or private university, access to an
education of quality (private?) to resources of the Information and Communication
Technology era, such as computers with Internet connection; and, beyond the current
income, the permanent income based on fixed socio-demographic features (like gender,
age, region, etc. but especially education stock). Concerning the sustainable aspect of
the family financial situation in the longer term, there is access to formal jobs that
ensure a higher level of social protection, access to private pension, to housing credit,
lease or freehold (with a minimum standard of quality: toilettes, type of building, etc.),
health insurance. This kind of concern with education and work placement is present in
the criteria applied in England, Portugal and India. The innovating aspect of this
methodology is its capacity to look at symbolic aspects of the middle class, such as the
employment registration booklet, university entry, or access to ICT, as well as to
combine these with social status aspects related to the private demand for goods, which
were previously a quasi monopoly of the Brazilian State, such as social security, health,
education and housing credit. Another innovation is the capacity to measure on a
national scale each mentioned component, to study its interaction and the aggregation of
the same ones into synthetic indexes about the size and distribution of the middle class;
as well as to look deeper into the details of its determining factors (for instance, to go
beyond the statistics concerning access to education to see how much it costs) and to
consider the interaction of these diverse components, monitoring them through time. In
the present research, we explore some aspects that are associated with the two lines
above mentioned: generation of income from work and labor mobility.
3
These variables are measured by the demographic Census, which facilitates the spatial classification of the families’
purchasing Power, but it is not well covered by the PNAD, for instance. In the hierarchical model that imputes
lacking income, as developed by IBGE, considered in the Census 2000 microdata, the most relevant variables are the
level of instruction of the person taken as a reference for the household and the number of toilettes
21
QUANTIFYING THE SIZE OF SOCIAL STRATA
Contrary to the analysis of relative income distribution where we map the relative share
of each group in the total income (like, for instance, the 10% richest who own almost
50% of the income in Brazil, etc.), here we focus on the share of population which is
between fixed thresholds for the whole period. In other words, we are concerned with
the absolute income of each person. The current approach is similar to that used in the
analysis of absolute poverty, only that we are concerned also with other frontiers, such
as those that determine the entry into the middle class and the upward movement
towards the upper-class. In the relative approach, the sum of the parts totals 100% of
something relative per month, while in the absolute approach applied to the several
segments in the social pyramid, they are indexed by an absolute value applicable to all
months. These absolute values are parameters for what it is to live in poverty, in an
intermediate group between poverty and the middle class, which we call here the D
class – (“remediados” or remediated), C class (middle class) and the A and b Classes
(upper-class). As our work concerns a period of strong growth in average income, the
two approaches - relative and absolute - present truly distinguished results. Each one of
these situations tend to happen, respectively, at the beginning and at the end of a period.
To use an analogy, in the relative distributive analysis, we may use a graph of a fixedsize pie, where for a group to win, another must loose its share. In the absolute analysis,
used here, apart from the distributive dance, the size of the pie can also change. What
lies behind the result is that, apart from people with lower income having gained a
greater share of the pie (reduction in inequality), the same cake has increased its size
(growth). The pie has grown to a medium-sized format; but for the optimists, this is now
a large cake. In the present analysis, we are not only concerned with its distribution, but
also with the amount of cake owned by the strata in society.
22
5. CHANGES IN THE PIE
The main feature of the approach used here is its level of disaggregation along three
income groups, where we look at the initial relative position first in order to probe them
further later on: the upper-class (the tenth richest part who owns almost half of the per
capita income); the poorest half how owns a little more than a tenth of the national
income (9,95%) including also the extremely poor; and the 40% intermediate group
whose share in the population and income virtually coincides with one another
(39,78%), making Brazil a medium income country, similar to Peru, and between the
rich Belgium and the poor India4.
Heuristically, here we want to investigate the
migrations between these different Brazils, as we are concerned about quantifying the
population strata living in the pre-fixed and determined living conditions and their
evolution through time.
The tables and graphs below present the levels and trends of these series.
TABLE 6 – Economic Classes
Labor Earnings Distribution – 15 to 60 years old
(6 Brazilian metropolitan areas)
Level
Apr/02
Apr/03
Apr/04
Apr/05
Apr/06
Apr/07
Apr/08
Middle Class
C Class
Moving
average
Rate (%) Rate (%)
44,19
42,49
43,42
42,26
42,80
46,70
46,21
48,59
48,72
48,87
50,11
51,89
50,81
Upper Class
A & B Class
Moving
average
Rate (%) Rate (%)
12,99
11,59
12,97
11,61
11,55
12,61
12,57
13,60
13,20
14,41
14,26
15,52
15,19
Poor & Just above
Poverty
E & D Class
Moving
average
Rate (%) Rate (%)
42,82
45,92
43,61
46,13
45,65
40,70
41,23
37,80
38,09
36,73
35,64
32,59
33,99
Just above Poverty
D Class
Moving
average
Rate (%) Rate (%)
14,18
15,46
14,44
15,68
15,40
15,28
15,24
13,25
14,00
15,01
13,71
14,20
14,22
Obs: 12-month moving average ended in the period
4
Under this aspect, the distribution of labor earnings in metropolitan areas in the PME is more
concentrated than in the national Pnad in all income sources.
23
Class C – New Middle Class
Firstly and more importantly for the objectives of this study, is the middle class. This
group encompassed 44.19% of the population at the beginning of the series in April
2002, moved on to 51.89% in April 2008 (the last observation available), with an
overall increase of 17.03% in the importance of the middle-class. If we fix the initial
period for after the 2002 instability and the 2003 recession, straight into April 2004, the
middle-class reached 42.85% of the population and grew around 18.72% until April
2008.
That is, a growth of almost 4% per year above the population growth of the reference
group.
TABLE 7 - Share in Income – 15 to 60 years old – Household Labor
(6 Brazilian metropolitan areas)
C Class
Moving
average
Variation (%) Variation (%) difference (p.p.)
Annual Evolution (12 months)
April 03 / April 02
April 04 / April 03
April 05 / April 04
April 06 / April 05
April 07 / April 06
April 08 / April 07
-3,85
-0,54
10,50
4,06
0,57
6,18
-1,42
7,96
5,43
2,85
1,41
-1,70
-0,23
4,44
1,90
0,28
3,02
Accumulated Evolution (since 2002)
April 08 / April 02
April 08 / April 03
April 08 / April 04
April 08 / April 05
April 08 / April 06
April 08 / April 07
17,43
22,13
22,79
11,12
6,79
6,18
17,03
18,72
9,97
4,30
1,41
24
7,70
9,40
9,63
5,19
3,30
3,02
GRAPH 7 - Participation of C Class (%) - Middle Class
Class structure - 15 to 60 years old - Labor – 6 metropolitan regions (Brazil)
52
50
48
46
44
Source: CPS/FGV based on PME microdata
Criterion: Per Capita Usual income (40% intermediate PME)
GRAPH 8 - Participation of C Class (%) - Middle Class
Class structure - 15 to 60 years old
12-month moving average
52
50
48
46
44
Source: CPS/FGV based on PME microdata
Criterion: Per Capita Usual income (40% intermediate PME)
25
Feb-08
Nov-07
Aug-07
May-07
Feb-07
Nov-06
Aug-06
May-06
Feb-06
Nov-05
Aug-05
May-05
Feb-05
Nov-04
Aug-04
May-04
Feb-04
Nov-03
Aug-03
May-03
Feb-03
42
Mar-08
Dec-07
Sep-07
Jun-07
Mar-07
Dec-06
Jun-06
Sep-06
Mar-06
Dec-05
Jun-05
Sep-05
Mar-05
Dec-04
Jun-04
Sep-04
Mar-04
Dec-03
Jun-03
Sep-03
Mar-03
Dec-02
Jun-02
Sep-02
Mar-02
42
TABLE 8 - Share in Income – 15 to 60 years old – Household - Labor
A and B Classes
Evolution
A & Br
Classes
Moving average
Variation
(%)
Variation (%)
-10,78
0,19
-10,95
8,57
8,82
7,89
5,01
5,90
8,02
7,73
6,56
April 03 / April 02
April 04 / April 03
April 05 / April 04
April 06 / April 05
April 07 / April 06
April 08 / April 07
April 08 / April 02
April 08 / April 03
April 08 / April 04
April 08 / April 05
April 08 / April 06
April 08 / April 07
19,46
33,89
33,64
23,09
14,09
7,73
Difference (p.p.)
-1,40
0,02
0,99
1,00
0,80
1,11
2,53
3,93
3,91
2,91
1,92
1,11
17,13
31,53
20,87
15,11
6,56
Back to the upper class group, which corresponds to classes A and B of the studies of
consumption potential, it reached 12.99% of the population at the beginning of the
series in April 2002, moved on to 15,52% in April 2008, with an increase of 19.46% in
importance reflecting the good period of the middle class.
GRAPH 9 - Participation of A & B Class (%)
Class structure - 15 to 60 years old - Labor – 6 metropolitan regions (Brazil)
16
15
14
13
12
Source: CPS/FGV based on PME microdata
Criterion: Per Capita Usual income
26
Mar-08
Dec-07
Sep-07
Jun-07
Mar-07
Dec-06
Jun-06
Sep-06
Mar-06
Dec-05
Jun-05
Sep-05
Mar-05
Dec-04
Jun-04
Sep-04
Mar-04
Dec-03
Jun-03
Sep-03
Mar-03
Dec-02
Jun-02
Sep-02
Mar-02
11
GRAPH 10 - Participation of A & B Class (%)
Class structure - 15 to 60 years old - 12-month moving average
16
15
14
13
12
Feb-08
Nov-07
Aug-07
May-07
Feb-07
Nov-06
Aug-06
May-06
Feb-06
Nov-05
Aug-05
May-05
Feb-05
Nov-04
Aug-04
May-04
Feb-04
Nov-03
Aug-03
May-03
Feb-03
11
Source: CPS/FGV based on PME microdata
Criterion: Per Capita Usual income
TABLE 9 - D & E Classes
Moving
average
April 03 / April 02
April 04 / April 03
April 05 / April 04
April 06 / April 05
April 07 / April 06
April 08 / April 07
April 08 / April 02
April 08 / April 03
April 08 / April 04
April 08 / April 05
April 08 / April 06
April 08 / April 07
Variation
(%)
Variation (%) Difference (p.p.)
7,24
3,10
0,45
4,67
0,21
-11,78
-9,69
-5,43
-7,11
-7,62
-2,89
-2,85
-6,43
-1,08
-11,26
-4,60
-4,14
-23,89
-29,03
-29,35
-19,92
-13,79
-11,26
-22,05
-25,53
-17,54
-10,74
-4,60
-10,23
-13,33
-13,54
-8,10
-5,21
-4,14
Criterion: Per Capita Usual income
Source: CPS/FGV based on PME microdata
Given our focus on the middle class and how we have already focused on the extremely
poor group in the section of the paper using the total population as reference, we restrict
this section to the sum of extremely poor and ´Just above Poverty´ groups, and then we
present the data for the D Class. The basis of the distribution formed by these two
groups corresponded to 42.82% of the population in April 2002, falling to 32,59%.
27
GRAPH 11 - Participation of D & E Classes (%)
Class structure - 15 to 60 years old - Labor – 6 metropolitan regions (Brazil)
46
44
42
40
38
36
34
Source: CPS/FGV based on PME microdata
GRAPH 12 - Participation of D & E Classes (%)
Class structure - 15 to 60 years old 12-month moving average
46
44
42
40
38
36
34
Source: CPS/FGV based on PME microdata
TABLE 10 - D Class (only)
April 03 / April 02
April 04 / April 03
April 05 / April 04
April 06 / April 05
April 07 / April 06
April 08 / April 07
Moving
average
Variation (%) Variation (%) Difference (p.p.)
9,05
1,28
1,40
6,61
0,22
-2,58
-1,05
-0,41
-13,24
-8,08
-2,02
13,26
-2,08
1,76
-5,42
3,68
-0,81
April 08 / April 02
0,11
April 08 / April 03
-8,20
-1,56
April 08 / April 04
-9,47
-7,67
April 08 / April 05
-7,07
-6,69
7,12
1,52
April 08 / April 06
April 08 / April 07
-5,42
3,68
Source: CPS/FGV based on PME microdata
28
0,02
-1,27
-1,48
-1,08
0,94
-0,81
Feb-08
Nov-07
Aug-07
May-07
Feb-07
Nov-06
Aug-06
May-06
Feb-06
Nov-05
Aug-05
May-05
Feb-05
Nov-04
Aug-04
May-04
Feb-04
Nov-03
Aug-03
May-03
Feb-03
32
Mar-08
Dec-07
Sep-07
Jun-07
Mar-07
Dec-06
Jun-06
Sep-06
Mar-06
Dec-05
Jun-05
Sep-05
Mar-05
Dec-04
Jun-04
Sep-04
Mar-04
Dec-03
Jun-03
Sep-03
Mar-03
Dec-02
Jun-02
Sep-02
Mar-02
32
As we separate them? from the analysis, we note a more erratic behavior of the D Class
through time inversely reflecting the middle-class trajectory, the other intermediate
group analyzed here. As a consequence, the participation of the group in the population
remains constant in comparison with the extremes in the series: goes from 14.18% in
april 2002 to 14.2% in April 2008.
GRAPH 13 - Participation of D Class (only) (%) Class structure - 15 to 60 years old
Labor – 6 metropolitan regions (Brazil)
16,0
15,5
15,0
14,5
14,0
13,5
13,0
Mar-08
Dec-07
Sep-07
Jun-07
Mar-07
Dec-06
Sep-06
Jun-06
Mar-06
Dec-05
Jun-05
Sep-05
Mar-05
Dec-04
Jun-04
Sep-04
Mar-04
Dec-03
Jun-03
Sep-03
Mar-03
Dec-02
Jun-02
Sep-02
Mar-02
12,5
Source: CPS/FGV based on PME microdata
GRAPH 14 - Participation of D Class (only) (%)
Class structure - 15 to 60 years old 12-month moving average
16
16
15
15
14
14
13
Source: CPS/FGV based on PME microdata
29
Feb-08
Nov-07
Aug-07
May-07
Feb-07
Nov-06
Aug-06
May-06
Feb-06
Nov-05
Aug-05
May-05
Feb-05
Nov-04
Aug-04
May-04
Feb-04
Nov-03
Aug-03
May-03
Feb-03
13
6. THE DISTRIBUTIVE DANCE
PME uses the methodology of a revolving panel that seeks to collect information in the
same households in the months t , t+1 , t+2 , t+3 , t+12 , t+13 , t+14 , t+15, totaling 8
interviews along a period of 16 months. The initial approach used here consists in
calculating the likelihood of the transition into- and out of poverty of the four society
groups such as the middle class, D Class, and upper class, as well as the non-transition
among these groups in the consecutive 4 months from March 2002. The last analyzed
group covers the period between January and April 2008.
The longitudinal aspect of the data for the household per capita income from work will
provide empirical evidence about the pattern of social mobility observed in practice. To
begin with, we present the data concerning people who enter or leave the two main
statuses through time, namely the middle-class (14.87%) and poverty (11.34%). This
means that at each four-month period from 2002 to 2008, 14.87% of the people entered
or left the middle-class and another 11.34% entered or left poverty. The greatest
mobility in relation to the middle-class and from it towards the upper groups, as well as
the people who leave the upper strata towards the middle-class and from it to the lowest
strata. Now, entry and exit out of poverty only happen for upper groups, causing a lower
likelihood of transition in relation to the movements related to the middle-class, with a
total of four alternatives of changes. The table below details the mobility rate relative to
each social group and the probability of associated transitions.
In four years, the time series indicate that the poverty rate based on income from work
fell 19.8% or 4.6 percentage points, from 23.3% to 18.57%. The biggest part of the
decrease happened between April and July 2004. When working with a moving average,
it becomes clear that the change in the poverty threshold occurred exactly in this period.
We will initially focus on this period, taking advantage of the PME longitudinal aspect.
Longitudinal data allow to measure who gets into and out of poverty, in the same way
as who remains in it or out of it through time. Table 9 analyzes the four-month period
between April and July each year.
30
TABLE 11
E Class
Mobility
Rate
Middle
Class
Mobility
Rate
E Class Exit
Probability
E Class
Probability
of entering
C Class Exit
probabiltiy
C Class
Probability
of entering
11,34
14,87
24,47%
6,88%
14,97%
14,76%
Total
Next, we present the evolution of these mobility rates through time. These series show that,
after the 2002 instability, there is a decrease in both the mobility rates between several society
groups until December 2005; then the series present some fluctuation with a marked upward
trend, especially in the middle-class mobility at the end of 2007 onwards. These data may be
capturing another type of risk in the case of income that affects the daily life of people. The
mobility peaks at the end of 2002 and since the 2007 coincide with the increased
macroeconomic uncertain conditions.
GRAPH 15 - Labor Mobility Rate (2002 a 2008)
6 Metropolitan Areas
21
19
17
15
13
11
9
E Class
Feb-08
Oct-07
Jun-07
Feb-07
Oct-06
Jun-06
Feb-06
Oct-05
Jun-05
Feb-05
Oct-04
Jun-04
Feb-04
Oct-03
Jun-03
Feb-03
Oct-02
Jun-02
7
C Class
Source: CPS/IBRE/FGV based on PME/IBGE microdata
In the analysis of the relative social mobility between the Rio de Janeiro and Salvador
Metropolitan areas, they are presented as less risky in relation to both segments, while the
riskier ones differ. Belo Horizonte and Porto Alegre are singled out in the total probability of
entering or leaving the middle-class, while Recife is singled out in terms of movements towards
poverty.
31
GRAPH 16 - Labor Mobility Rate (2002 a 2008)
by Metropolitan Areas
19,5
17,8
16,8
19,3
16,1
14,9
13,9
10,7
E Class
Porto
Alegre
São Paulo
Rio de
Janeiro
Belo
Horizonte
8,5
Salvador
Recife
10,3 10,3
11,0
C Class
Source: CPS/IBRE/FGV based on PME/IBGE microdata
As we isolate the entry and exit movements of each status, we gain a clearer vision of
their nature. The risks of entering poverty have fallen slightly through time, while the
risks of leaving it not only have reached higher levels, but have also fluctuated a little
more. It is worth highlighting the two peaks in the series of poverty exit in mid-2004
recovery after the 2003 recession and the positive record observed in 2008. That is,
according to the Chinese hexagram that depicts the junction of two trigrams of risk and
opportunity, 2007 represents paradoxically a rise in the risk of finding good
opportunities.
GRAPH 17 - Labor Mobility Probability (2002 a 2008) – E Class
6 Metropolitan Areas
35%
30%
25%
20%
15%
10%
5%
Probability of exit
Probability of entering
Source: CPS/IBRE/FGV based on PME/IBGE microdata
32
Feb-08
Oct-07
Jun-07
Feb-07
Oct-06
Jun-06
Feb-06
Oct-05
Jun-05
Feb-05
Oct-04
Jun-04
Feb-04
Oct-03
Jun-03
Feb-03
Oct-02
Jun-02
0%
Now, we move back to the open regional data according to the entry and exit of the
income lower end. The risks faced by Recife inhabitants in entering and leaving poverty
in the period are very close. In particular, the probability of entering poverty, although
lower than entry’s, is around three times as high if compared to Rio and more than 60%
higher than Belo Horizonte’s, which has the second highest probability of poverty
entry. On a positive note, the probability of poverty exit in Porto Alegre, Belo Horizonte
and São Paulo are in much higher stances compared to the three other areas.
GRAPH 18 - Labor Mobility Probability (2002 a 2008) – E Class
By Metropolitan Areas
29,4%
20,4%
19,1%
17,1%
9,4%
Probability of exit
Rio de
Janeiro
6,3%
São Paulo
4,7%
Belo
Horizonte
Salvador
6,8%
Recife
30,8%
8,7%
Porto
Alegre
15,8%
29,4%
Probability of entering
Source: CPS/IBRE/FGV based on PME/IBGE microdata
Next, we present the probability of entering and leaving the middle-class by
metropolitan region, which for its bilateral aspect is more complex to be analyzed than
that of poverty, where it is clearer that he who enters poverty or leaves it may,
respectively, lose or gain purchasing power. Briefly, Recife is highlighted for its greater
probability of entering the middle-class; whereas Porto Alegre has the highest
probability of exit.
33
GRAPH 19 - Labor Mobility Probability (2002 a 2008) – C Class
By Metropolitan Areas
27,3%
19,6%
19,4%
13,6%
Probability of exit
Porto
Alegre
Rio de
Janeiro
Belo
Horizonte
Salvador
Recife
20,6%
18,0%
11,4%
9,9%
8,4%
São Paulo
12,2%
17,3%
15,0%
Probability of entering
Source: CPS/IBRE/FGV based on PME/IBGE microdata
SOCIAL DESTINATIONS
Finally, we uncover the destinations for each social strata transition per year. Once
more, 2004 and 2008 are remarkable in terms of poverty; only 71.96% of the poor
remained poor four months after the first observation in 2004. Such statistics fall to
67.57% in the period observed in 2008 – which saw the transitions from Class E ( poor)
towards classes D and C. That is, we analyzed the transition into and out of the different
income groups. In all cases, 2008 is the best point in the series. The reader is invited to
look at the destinations of the individuals from several social segments.
TABLE 12 - Destination Matrix - (Who was Class E initially)
Period2 (April)
E Class
D Class
C Class
A&B
classes
January
2003
73,38
13,48
11,30
1,85
2004
71,96
14,13
11,52
2,38
2005
77,30
13,20
8,87
0,63
2006
76,86
12,65
9,66
0,83
2007
77,52
11,22
10,22
1,03
2008
67,57
16,46
14,53
1,45
Source: CPS/IBRE/FGV based on PME/IBGE microdata
34
TABLE 13 - Destination Matrix - (Who was Class D initially)
Period2 (April)
E Class
D Class
C Class
A&B
classes
January
2003
53,30
23,66
22,72
0,33
2004
56,19
19,13
24,27
0,41
2005
62,28
16,03
21,56
0,13
2006
58,13
18,58
22,89
0,41
2007
65,00
14,75
20,26
0,00
2008
54,92
15,07
29,46
0,56
Source: CPS/IBRE/FGV based on PME/IBGE microdata
TABLE 14 - Destination Matrix - (Who was middle-class (Class C) initially)
Period 2 (April)
E Class
D Class
C Class
A&B
classes
January
2003
78,78
8,82
10,02
2,39
2004
81,04
6,26
8,55
4,15
2005
86,95
4,20
6,35
2,50
2006
85,11
5,63
6,17
3,10
2007
86,35
3,76
6,52
3,37
2008
84,58
3,70
6,98
4,74
Source: CPS/IBRE/FGV based on PME/IBGE microdata
TABLE 15 - Destination Matrix - (Who was in the upper class initially)
Period 2 (April)
E Class
D Class
C Class
A&B
classes
January
2003
78,04
4,50
0,39
17,07
2004
77,64
4,68
0,74
16,95
2005
89,11
1,32
0,82
8,75
2006
85,51
1,96
0,19
12,34
2007
86,91
1,00
0,14
11,95
2008
82,38
1,60
0,55
15,47
Source: CPS/IBRE/FGV based on PME/IBGE microdata
35
7. THE RETURN OF THE “CARTEIRA DE TRABALHO” (EMPLOYMENT
REGISTRATION BOOKLET)
What is it to be part of the Class C? Having a computer, a car, credit access, a mortgage,
a mobile, being self-employed or an employer, making pension payments; having a
university degree, attending a private school, health insurance. But of all, the return of
the employment registration booklet may be the most representative element of the new
Brazilian middle class appearance. The final step was to analyze the evolution of formal
job posts in the country. This information is particularly important, since having a job
with a registration is a strong feature of the middle class. In this context, the recent
information is encouraging, with 309 thousand jobs in just a month, a record in the
historical series in June 2008 and 1.881 million new formal job posts in the last 12
months.
GRAPH 20 - Net Job Creation
300.000
200.000
100.000
06_2008
11_2007
04_2007
09_2006
02_2006
07_2005
12_2004
05_2004
10_2003
03_2003
08_2002
01_2002
06_2001
11_2000
04_2000
09_1999
02_1999
07_1998
12_1997
05_1997
10_1996
03_1996
0
-100.000
-200.000
-300.000
Source: CPS/IBRE/FGV based on CAGED/ M T E.
jun(2008)
2
1.88
1 .09
jun(2007)
1.21
0 .19
2
1.45
4 .11
1.39
8 .83
0
564
.0 69
jun(2002)
642
.9 08
531
.4 96
jun(2001)
jun(2006)
jun(2005)
jun(2004)
jun(2003)
jun(2000)
-624
.600
jun(1999)
-295
.269
jun(1998)
jun(1997)
-45.
617
376
.3 25
1.11
6 .73
0
5
GRAPH 21 - Net Job Creation ( 12-month accumulated)
Source: CPS/IBRE/FGV based on CAGED/ M T E.
36
Next, we zoom into the 6 metropolitan areas in Brazil. The proportion of jobs created
there has grown, reaching 387 thousand posts in the first semester of 2008, around
28.5% of the total – the highest proportion in a historical series since 1992. In 2003, this
percentage was only 12.45%.
GRAPH 22 - Net Job creation (12-month accumulated)
1125382
956829
858359
804250
729148
738461
493672
489641
6 RM s PM E
jun(2008)
jun(2007)
jun(2006)
jun(2005)
298768
jun(2004)
jun(2003)
145437
539901
491414
Other Regio ns
Source: CPS/IBRE/FGV based on CAGED/ M T E.
In the next tables, we present a overview of the evolution in the different metropolitan
areas.
TABLE 16 - Net Job Creation
jan to june each year
2008
2007
2006
2005
2004
2003
2002
2001
2000
1999
1998
1997
1996
Total
TOTAL
Other
regions
TOTAL
RM PME
RECIFE
SALVADOR
BELO
HORIZON
R DE
JANEIRO
SAO
PAULO
PORTO
ALEGRE
1.360.645
1.095.249
923.937
965.695
1.033.289
561.866
680.443
573.544
589.796
18.055
61.267
321.989
95.547
8.281.322
973.458
815.172
695.830
716.473
814.786
491.921
567.642
443.050
448.892
104.921
132.495
292.842
128.475
6.625.957
387.187
280.077
228.107
249.222
218.503
69.945
112.801
130.494
140.904
-86.866
-71.228
29.147
-32.928
1.655.365
7.371
2.703
809
3.362
1.305
-6.894
-553
1.754
2.228
-9.660
-9.190
-8.831
-10.005
-25.601
19.505
13.474
5.400
13.637
6.900
-1.777
7.762
4.728
6.809
-6.638
-4.504
4.592
-3.983
65.905
57.108
38.347
40.190
36.944
30.037
9.134
17.113
14.026
17.470
-3.469
-2.353
16.710
8.128
279.385
56.896
44.949
33.615
27.189
30.708
7.288
15.365
14.553
20.235
-19.719
-5.862
-3.436
4.145
225.926
217.076
163.913
138.280
157.178
121.136
50.627
66.370
80.917
73.180
-42.559
-41.926
13.740
-22.465
975.467
29.231
16.691
9.813
10.912
28.417
11.567
6.744
14.516
20.982
-4.821
-7.393
6.372
-8.748
134.283
3.639.847
3.397.757
4.960.258
Inhabitants 183.305.600 135.868.370 47.437.230
Source: CPS/IBRE/FGV based on CAGED/ M T E.
37
11.682.332 19.666.573 4.090.463
TABLE 17 - Net Job Creation (% Total)
jan a junho de cada ano
Other
regions
TOTAL
RM PME
RECIFE
SALVADOR
BELO
HORIZON
R DE
JANEIRO
SAO
PAULO
PORTO
ALEGRE
2008
2007
2006
2005
2004
2003
2002
2001
2000
1999
1998
1997
1996
Total
71,54
74,43
75,31
74,19
78,85
87,55
83,42
77,25
76,11
581,12
216,26
90,95
134,46
80,01
28,46
25,57
24,69
25,81
21,15
12,45
16,58
22,75
23,89
-481,12
-116,26
9,05
-34,46
19,99
0,54
0,25
0,09
0,35
0,13
-1,23
-0,08
0,31
0,38
-53,50
-15,00
-2,74
-10,47
-0,31
1,43
1,23
0,58
1,41
0,67
-0,32
1,14
0,82
1,15
-36,77
-7,35
1,43
-4,17
0,80
4,20
3,50
4,35
3,83
2,91
1,63
2,51
2,45
2,96
-19,21
-3,84
5,19
8,51
3,37
4,18
4,10
3,64
2,82
2,97
1,30
2,26
2,54
3,43
-109,22
-9,57
-1,07
4,34
2,73
15,95
14,97
14,97
16,28
11,72
9,01
9,75
14,11
12,41
-235,72
-68,43
4,27
-23,51
11,78
2,15
1,52
1,06
1,13
2,75
2,06
0,99
2,53
3,56
-26,70
-12,07
1,98
-9,16
1,62
Inhabitants
74,12
25,88
1,99
1,85
2,71
6,37
10,73
2,23
Source: CPS/IBRE/FGV based on CAGED/ MTE.
38
8. CONCLUSIONS
This study was based on the PME microdata for their agile nature, which has enabled us
to diminish the average PNAD time lag of 18 months down to only four months. This
allowed us to assess the trajectory of the social indicators up to mid-2008. The increase
in speed is a necessary condition to foresee the continuity, or not, of the improvement in
distribution observed since the beginning of the decade. We observe the sustainability
of the pace of inequality reduction in the income from work between the end of 2006
and mid-2008. If the reduction in poverty were a competition among the workers of the
6 metropolitan regions in Brazil, these would be the award winning moments in the last
6 years: firstly the last 12 months: 2007-08 (-13,5%) followed by the same period in
2004-05 (-12,3%) and then 2006-07 (-8%). It is worth remembering that the last and the
first have not yet been considered by the official social statistics.
We focused on the greatest metropolitan areas as a result of their great availability of
recent data (until 2008). The core point concerns the finding that after years of crises in
these areas, the growth of a middle class there and how much of it is attributable to the
private job creation. The emphasis of the analysis is on the 6 biggest Brazilian
metropolitan areas that show a recovery from the so-called metropolitan unemployment
crisis. The research also emphasizes the social performance in these cities taken
isolately. In many cases, the analyses focus on the role of the high public income
transfers to the population such as the expansion of the Bolsa Familia and of the social
security contribution-related transfers associated with minimum wage adjustments. We
argue that, at least, the increase in the income from work rivals with these transfer in
explaining improvements in the income for the whole population (and that since 2001
for the poorest groups). This is well documented in the PNAD series that go as far as
October 2006. But since the end of the 2006 until now, the unprecedented aspect of the
empirical analysis, has been the leadership role of income from work in general and of
formal job posts created, in particular. That is, since the last statistical snapshot taken of
Brazil through the 2006 Pnad lenses, income from work has been the most important
aspect.
In effect, there is a lack of tune between the PNAD income and that from the national
accounts in the period since 2004 (Neri 2007 and 2005). In the period of 2005-06,
PNAD income grew at a Chinese-like rhythm of 16.4% 4.3 times higher than the Haiti-
39
like speed of per capita GDP. Pnad has not undergone any methodological change, nor
did the INPC used to deflate this calculation. The Asian looks of the PNAD growth is
reflected in other indicators in the 2005-06 period, such as commerce sales 11.8%, the
valuation of the stock markets, Ibovespa increased at 60%, and the creation of 4.6
million job posts, in particular, 2.5 million of new formal posts. Since 2001, Brazil
experiences a Chinese-like growth for the poorest (and only for them) but in 2005 and
2006 it experienced a Chinese-like growth for all social strata, compared to 2004. We
bring this discussion to 2007 and 2008 and afterwards.
In general terms, data point to a sustained decrease in poverty and an expansion of the
so-called middle class observed since the end of the 2003 recession. The rhythm in the
reduction of inequality observed since 2001, has not slowed down, and can be compared
in absolute terms with the famous concentration of income in the 1960’s, time of the
Brazilian economic miracle. The increase in average income keeps the pace of the
previous years as a result of the previous period, despite the slower rhythm observed in
the central economies and in the USA in particular. In short, the cake keeps growing
with more strength for the poorest classes for over 5 years, an unprecedented
combination in the Brazilian documented statistical history. Much has been said in the
decade about reduction in inequality and poverty, but much emphasis has been given to
the role of public income transfers to the poor and little attention to the structural
advances of the remaining segments in society.
The research reveals the recent rise of a new middle class despite the approaching signs
of external crises from the USA. Overall, the new PME data allow us to monitor the
social performance of different segments in the 6 metropolitan areas. After several years
of crises in these areas, they are back in the scene. The regional aspect which draws
more attention is the better performance of the Belo Horizonte area. Now, a question
related to the election period: which of the Brazilian metropolitan areas have had the
best reduction of poverty in the last six years? The answer would be Belo Horizonte (40,8%) , Rio de Janeiro (-30,7%) followed closed by Salvador (-29,8%). The
movements of improvements in the series of social indicators in the Great Belo
Horizonte region are relatively dispersed through 2002 to 2008, while the remaining
metropolitan areas have seen it more between April and June 2004. This geographical
and time aspects aside, a reduction in the distance between the average and the median
40
is present in all regions, representing the reduction in inequality that marks the recent
period.
Even in the period covered by the Pnads, the PME data provide revealing details of their
determinants through an analysis of the monthly dynamics of the indicators based on
income. In particular, we note a marked growth in the incomes of the poorest classes
between March and June 2004. Although this is the time when the minimum wage is
readjusted, there was not a real gain in the minimum wage during 2004. PME allows
assessing the details of entry and exit of several segments of the population, as it
follows the trajectory of the same individuals and their families through time.
To complete this recent scenario, we also counted on the Caged/MTE data (Cadastro
Geral de Emprego e Desemprego do Ministério do Trabalho e Emprego) from all
national territory. In this context, recent information is encouraging, with 309 thousand
jobs in a month reaching a record in the historical series now in June 2008 and 1,881
million of new formal job posts in the past 12 months. Then we zoomed into the 6 main
metropolitan areas. The proportion of jobs created there has grown, with 387 thousand
posts in the first semester of 2008, or 28.5% of the total, the biggest proportion in the
historical series (since 1992). In 2003, this was only 12.45%.
We begin this research from two perspectives about class definitions. Firstly, the
analysis of people’s attitudes and expectations. This kind of approach was developed in
the 1950s and 1960s’ by George Katona, behavioral psychologist. Following this line,
Thomas Friedman defines middle-class as having a well-defined social ascension plan
for the future. CPS proposes measurements in this subjective line: our Future Happiness
Index (IFF) will be released soon. We may already say that the index is high in Brazil
vis-à-vis other countries, which is consistent with the emergence of a middle-class in
the country.
The second way to define the economic classes E, D, C, B and A is according to their
consumption potential. The Brazil Criterion uses access to - and number of durable
goods (TV sets, radio, washing machine, fridge and freezer, video or DVD player),
toilettes, housemaid and the level of education of the head of the household. This
criterion estimates the weighs based on a Mincerian income equation. Beside measuring
41
all this, we suggest a complementary conceptualization to measure the evolution of the
new middle-class in Brazil also from the producer’s standpoint - in other words, based
on his/her capacity to keep this consumption potential through time. Since Robert
Hall’s 1977 classic work, we’ve known that the current consumption contains
information about the future patterns of consumption, whereas the analysis of these
expenses composition is useful by separating hedonism from the productive capacity, to
the same level of personal expenses. In the same terms as La Fontaine’s fable, one must
distinguish the ants from the consumerist grasshoppers.
The always tuned IBGE is now carrying out an excellent Family Budget Survey (POF)
and has announced the monitoring of family consumption expenses on a regular basis.
As we haven’t got there just yet, we implemented a methodology to explore IBGE’s
microdata. PME covers 40 thousand households per month in the six main metropolitan
areas in Brazil. Apart from its geographical limitations, the survey focuses exclusively
on income from work. Beside the sample size, the survey’s strength resides in its recent
aspect as PME brings us as far as April 2008; while POF, 2002/03 and PNAD,
September 2006. Income inequality follows the same decreasing trend observed since
2001. There is nothing statistically similar in the Brazilian documented history (since
1960). Another PME’s advantage is to measure changes in income levels from the same
families through time.
Class C is the central class, below A and B and above D and E. Brazil’s class C grew
from 42% in 2004 to 52% in 2008. According to the Pew Institute, 53% of the NorthAmericans consider themselves to be part of the middle class. The new Brazilian
criterion pointed to 43% of Brazilians in the Class C in 2005, close to the 2004’s 42%.
In order to quantify the levels, we calculated the household per capita income from
work and then expressed it in terms equivalent to the total household income from all
sources. The central C class has an income between R$ 1064 and R$ 4561 at today’s
prices from the São Paulo Metropolitan region. International studies vary in what
concerns the upper monthly limit of the middle-class income from US$ 6000 (World
Bank), US$ 500 (Goldman Sachs) to US$300 (Barnajee & Duflo, MIT). Ours is within
their limits; that vary greatly. Our class C is located immediately above the 50% poorest
and the 10% richest at the turn of the century. Illustratively, class C is limited by the
42
Indian border on one side and by the Belgian border on the other side in the so-called
Belíndia. (Fn Belíndia is an aphorism created by Edmar Bacha to illustrate the Brazilian
income inequality).
We do not hold anything against those who see Brazilian class C as the lower middle
class and class B as the upper middle class. We are not the USA but this is also a free
country. It is indeed more important to have a well-defined, consistent criterion; but the
middle class label can be shared by other strata. All in all, our class C gets society’s
average income, that is, it is the middle-class in the statistical sense. Given the Brazilian
inequality, the average income is high in relation to our median. In relation to the rest of
the world: 80% of people in the world live in countries with a per capita income lower
than Brazilians’. Now, those who think that class C’s income is low must awaken
because this is the closest image to the Brazilian society. The Brazilian elite who take
themselves to be part of the middle-class should search for the words Made in USA
behind their mirrors.
43
9. BIBLIOGRAPHY
BACHA, E. L., AND TAYLOR, L.. “Brazilian income distribution in the 1960s: Tacts’
model results and the controversy.” Journal of Development Studies, Vol. 14, Issue 3,
pages 271 – 297, 1978
BANERJEE, ABHIJIT V. AND DUFLO, ESTHER ”What is middle class about the
middle classes around the world?”December 2007.
BARROS, R.P. de; MENDONÇA, R. A evolução do bem-estar e da desigualdade no
Brasil desde 1960. Rio de Janeiro: IPEA, 1992. (Texto para discussão, nº. 286).
BARROS, R.P. de; HENRIQUES, R.; MENDONÇA, R. Desigualdade e pobreza no
Brasil: a estabilidade inaceitável. In: HENIQUES, R. (Ed.). Desigualdade e pobreza
no Brasil. Rio de Janeiro: IIPEA, 2000.
BARROS, R. P. Foguel, M. N. ULYSSEA G. (Orgs.). Desigualdade de Renda no
Brasil: uma análise da queda recente. Rio de Janeiro: IPEA, 2007.
BARROS, Ricardo Paes; MENDONÇA, Rosane; NERI, Marcelo C. The duration of
poverty spells. In: III Encontro Nacional de Estudos do Trabalho, ENABET, Anais...,
1996.
BARROS, Ricardo Paes; MENDONÇA, Rosane; NERI, Marcelo C. Pobreza e inflação
no Brasil: uma análise agregada. In: Economia Brasileira em Perspectiva 1996, Rio
de Janeiro: Ipea, 1996, v.2, p.401-420.
BIRDSALL, NANCY, CAROL GRAHAM, AND STEFANO PETTINATO “Stuck In
Tunnel: Is Globalization Mudding The Middle Class?” Brookings Institution, Center
on Social and Economic Dynamics WP No. 14, 2000.
BOOT, H. M. (1999) “Real Incomes of the British Middle Class, 1760-1850: The
Experience of Clerks at the East India Company”, The Economic History Review,
52(4), 638-668.
BONELLI, R.P. de; SEDLACEK, G.L. Distribuição de renda: evolução no último
quarto de século. In: SEDLACEK, G.L.; BARROS, R.P. de. Mercado de trabalho e
distribuição de renda: uma coletânea. Rio de Janeiro: IPEA, 1989. (Série Monográfica
35).
CARDOSO, E.; BARROS, R.; URANI, A. Inflation and unemployment as determinants
of inequality in Brazil: the 1980s, Chapter 5. In: DORNBUSCH, R.; EDWARDS, S.
(Eds.), Reform, recovery and growth: Latin America and the Middle-East, Chicago:
University of Chicago Press for the NBER, 1995.
DOEPKE, M.AND F. ZILIBOTTI (2005)”Social Class and the Spirit of Capitalism”,
Journal of the European Economic Association 3, 516-24.
DOEPKE, M AND F. ZILIBOTTI (2007)”Occupational Choice and the Spirit of
Capitalism”, NBER Working Paper.
44
ESTERLY, WILLIAM (2001) “The Middle Class Consensus and Economic
Development”, Journal of Economic Growth, 6(4), 317-335.
FERREIRA, F.; LANJOUW, P.; NERI, M. A Robust poverty profile for Brazil using
multiple data sources. Revista Brasileira de Economia 57 (1), p. 59-92, 2003.
FISHLOW, A. (1972): “Brazilian Size Distribution of Income”, American Economic
Association: Papers and Proceedings 1972, pp.391-402
Frederick, Jim (2002) “Thriving in the Middle Kingdom”, TIME Magazine, Nov 11.
GASPARINI, L. Different lives: inequality in Latin America the Caribbean, inequality
the state in Latin America the Caribbean World Bank LAC Flagship Report 2003.
Washington, D.C.: World Bank, 2003. Mimeografado.
HOFFMAN, R. A evolução da distribuição de renda no Brasil, entre pessoas e entre
famílias, 1979/86. In: SEDLACEK, G.; BARROS R.P. de. Mercado de trabalho e
distribuição de renda: uma coletânea. Rio de Janeiro: IPEA/Inpes, 1989.
HOFFMANN, R. As transferências não são a causa principal da redução da
desigualdade, Econômica 7, no.2, 335-341: Rio de Janeiro, Brazil, 2005.
IPEA. Sobre a queda recente da desigualdade no Brasil, 2006. (Nota técnica).
KAKWANI, N., SON, H. Measuring the Impact of price changes on poverty.
International Poverty Centre, Brasília, 2006. (Working paper # 33).
KAKWANI, N.; NERI, M.; SON, H. Linkages between pro-poor growth, social
programmes labour market: the recent brazilian experience. International Poverty
Centre, Brasilia, 2006a. (Working paper # 26).
______. Desigualdade e Crescimento: Ingredientes Trabalhistas em Desigualdade
de Renda no Brasil: uma análise da queda recente. Ricardo Paes de Barros, Miguel
Nathan Foguel, Gabriel Ulyssea (orgs), Rio de Janeiro, 2007. vide
http://www.fgv.br/cps/pesquisas/propobre/
Landes, David. The Wealth and Poverty of Nations. New York: Norton, 1998.
LANGONI, C. Distribuição da renda e desenvolvimento econômico do Brasil. Rio
de Janeiro: Fundação Getúlio Vargas (FGV), 3a edição 2005, 1973
MURPHY, KEVIN M., ANDREI SCHLEIFER AND ROBERT VISHNY (1989)
“Industrialization and the Big Push”, Journal of Political Economy, 97(5), 10031026.
NERI, Marcelo C.; CONSIDERA, Cláudio; PINTO, Alexandre. A evolução da pobreza
e da desigualdade brasileiras ao longo da década de 90. In: Revista Economia
Aplicada, Ano 3, v. 3, p.384-406, jul.-set. 1999.
45
NERI, Marcelo C. O reajuste do salário mínimo de maio de 1995. In: XIX ENCONTRO
BRASILEIRO DE ECONOMIA, SBE, Recife. Anais... dez. 1997, v. 2, p. 645-666.
NERI, Marcelo C.; CONSIDERA, Cláudio. Crescimento, desigualdade e pobreza: o
impacto da estabilização. In: Economia Brasileira em Perspectiva 1996, Rio de
Janeiro: Ipea, 1996, v.1, p. 49-82.
NERI, M. C. Diferentes histórias em diferentes cidades. In: REIS VELLOSO, J.P.;
CAVALCANTI, R. (Eds.). Soluções para a questão do emprego. Rio de Janeiro: José
Olimpio, 2000.
______. Eleições e Expanções, mimeo, vide http://www.fgv.br/cps/pesquisas/pp2/,
2006a.
______. Miséria em queda: mensuração, monitoramento e metas. mimeo Rio de
Janeiro: FGV, vide: http://www3.fgv.br/ibrecps/queda_da_miseria/inicio_q.htm . 2005.
______. Miséria, desigualdade e politicas de rendas: o Real do Lula, mimeo, Rio de
Janeiro: FGV, 2007.
______. Miséria, desigualdade e estabilidade in Desigualdade de Renda no Brasil:
uma análise da queda recente. Ricardo Paes de Barros, Miguel Nathan Foguel, Gabriel
Ulyssea (orgs), Rio de Janeiro, 2007a. see:
<http://www.fgv.br/cps/pesquisas/site_ret_port/
______. A Dinâmica da Redistribuição Trabalhista em Desigualdade de Renda no
Brasil: uma análise da queda recente. Ricardo Paes de Barros, Miguel Nathan Foguel,
Gabriel Ulyssea (orgs), Rio de Janeiro, 2007b.
______. (org) Microcrédito, o mistério nordestino e o Grameen brasileiro: perfil e
performance dos clientes do CrediAMIGO”, Editora da Fundação Getulio Vargas,
370pag, Rio de Janeiro, 2008
NERI, M. C.; CAMARGO, J. Distributive effects of Brazilian structural reforms. In:
BAUMANN, R. (Ed.). Brazil in the 1990s: a decade in transition, Palgrave.
Macmillan's Global Academic Publishing, UK, 2001.
NERI, M. C.; GIOVANNI F, Negócios nanicos, garantias e acesso a crédito in
Revista de Economia Contemporânea, Rio de Janeiro, v.9, n.3, pp 643-669, setemberdecember 2005.
RAMOS, Lauro; BRITO, M. O funcionamento do mercado de trabalho metropolitano
brasileiro no período 1991-2002: tendências, fatos estilizados e mudanças estruturais.
Boletim Mercado de Trabalho, Conjuntura e Análise, Rio de Janeiro: Ipea, nº 22, p.
31-47, nov. 2003.
ROCHA, S. Pobreza no Brasil: afinal do que se trata? Rio de Janeiro: Ed. FGV, 2003.
46
SOARES, S. “Análise de bem-estar e decomposição por fatores da queda na
desigualdade entre 1995 e 2004.” Econômica, v. 8, n. 1, p. 83-115. Rio de Janeiro,
2006.
WILSON, DOMINIC AND DRAGUSANU, RALUCA “The Expanding Middle: The
Exploding World Middle Class and Falling Global Inequality” - Goldman Sachs
Economic Research/Global Economics Paper nº 170, July 2008
47
APPENDIX I. PORTRAIT OF THE NEW MIDDLE CLASS
In this section, we present the evolution of the Brazilian middle-class according to
different socio-economic characteristics since 2002.
Gender
Income Groups – Labor (6 Brazilian metropolitan areas)
C Class – New Middle Class (%)
Gender
Men
Women
2008* /
2002
2002
2003
2004
Period
2005
2006
2007
2008
18,89
17,53
44,88
42,51
44,36
41,75
46,18
43,51
49,76
46,63
51,57
48,45
51,74
48,51
53,36
49,96
2008* /
2002
2002
2003
2004
Period
2005
2006
2007
2008
27,14
30,18
28,43
31,73
25,71
29,30
22,50
26,26
20,76
24,60
18,68
22,52
16,81
20,96
E Class (%)
Gender
Men
-38,07
Women
-30,55
Source: CPS/FGV based on PME microdata
Age
Income Groups – Labor (6 Brazilian metropolitan areas)
C Class– New Middle Class (%)
Age
15 to 19
20 to 24
25 to 29
30 to 35
36 to 39
40 to 44
45 to 49
50 to 54
55 to 59
2008* /
2002
2002
2003
2004
Period
2005
2006
2007
2008
20,01
19,33
21,72
17,17
16,26
18,82
14,90
16,35
20,33
40,42
46,72
45,19
43,53
43,85
44,31
45,08
42,65
37,95
38,56
45,92
44,83
42,16
42,29
43,68
45,79
43,32
38,96
40,24
47,93
46,72
43,93
43,81
45,58
47,62
45,11
40,48
43,41
51,68
49,90
47,80
47,56
48,88
49,69
48,64
43,49
46,40
53,84
52,03
49,23
49,52
50,98
50,88
49,78
44,07
47,01
54,93
52,57
48,81
49,53
51,11
50,96
48,80
43,49
48,51
55,75
55,01
51,00
50,99
52,65
51,79
49,62
45,67
2008* /
2002
2002
2003
2004
Period
2005
2006
2007
2008
29,13
20,76
19,67
21,74
21,83
20,94
20,76
22,85
29,85
26,42
18,38
17,70
19,57
20,02
18,98
18,76
21,33
27,96
25,07
17,01
15,15
17,23
18,89
17,37
17,41
19,65
26,33
E Class (%)
Age
15 to 19
20 to 24
25 to 29
30 to 35
36 to 39
40 to 44
45 to 49
50 to 54
55 to 59
-28,02
34,83 37,12 34,46 31,63
-37,02
27,01 28,68 26,20 23,16
-42,14
26,19 27,93 25,42 21,94
-37,20
27,44 28,78 26,61 23,42
-29,72
26,88 29,10 26,82 23,32
-35,81
27,06 28,65 25,30 22,32
-31,90
25,56 26,57 23,97 21,71
-34,32
29,92 29,73 27,37 24,05
-28,64
36,90 36,88 33,95 30,35
Source: CPS/FGV based on PME microdata
48
Education
Income Groups – Labor (6 Brazilian metropolitan areas)
C Class– New Middle Class (%)
Education (years of schooling)
Without instruction, less than 1
year
From 1 to 3 years of schooling
From 4 to 7 years of schooling
From 8 to 10 years of schooling
11 or more years of schooling
2008* /
2002
2002
2003
2004
Period
2005
2006
2007
2008
32,44
22,48
19,97
16,10
14,51
28,63
35,88
41,11
48,22
45,46
27,98
32,87
38,74
46,54
46,68
28,68
34,41
40,61
47,67
48,43
31,18
38,07
43,86
51,37
51,30
33,74
40,57
46,51
53,75
52,02
35,04
41,32
47,62
53,80
51,22
37,92
43,95
49,32
55,98
52,06
2008* /
2002
2002
2003
2004
Period
2005
2006
2007
2008
46,27
39,27
32,36
23,93
13,79
43,36
35,67
29,61
21,94
13,06
40,09
32,47
27,97
20,49
12,03
E Class (%)
Education (years of schooling)
Without instruction, less than 1
year
From 1 to 3 years of schooling
From 4 to 7 years of schooling
From 8 to 10 years of schooling
11 or more years of schooling
-23,08
52,12 52,07 51,01 48,74
-25,81
43,77 46,48 44,50 40,82
-25,23
37,40 39,32 37,00 34,60
-26,86
28,02 30,23 28,36 25,58
-34,06
18,24 19,67 17,37 14,49
Source: CPS/FGV based on PME microdata
Job post
Income Groups – Labor (6 Brazilian metropolitan areas)
C Class– New Middle Class (%)
Job post
Formal employees
Informal employees
Employees military
Self-employed
Employer
Non-paid worker
Unemployed
Inactive
2008* /
2002
2002
2003
2004
Period
2005
2006
2007
2008
13,69
17,12
-2,41
18,77
8,58
10,86
31,98
14,72
56,51
49,24
50,78
47,04
35,85
49,91
28,49
33,21
56,66
48,09
50,89
46,45
38,61
47,71
27,20
31,63
58,48
51,03
52,71
47,97
40,67
49,18
28,66
32,51
62,20
53,57
53,99
51,20
41,55
57,00
30,55
35,45
63,48
56,18
52,38
53,63
41,32
52,84
33,32
36,93
63,23
55,99
49,97
53,88
39,37
52,19
34,33
36,79
64,25
57,66
49,55
55,87
38,93
55,33
37,60
38,11
2008* /
2002
2002
2003
2004
Period
2005
2006
2007
2008
7,39
15,58
5,50
17,75
5,71
20,68
46,15
40,50
6,42
13,85
5,25
15,24
6,94
17,96
42,87
38,21
5,60
12,59
4,36
13,29
5,37
14,78
39,99
36,46
E Class (%)
Job post
Formal employees
Informal employees
Employees military
Self-employed
Employer
Non-paid worker
Unemployed
Inactive
-55,11
12,48 13,54 11,05
8,64
-42,65
21,95 23,65 20,49 17,54
-54,68
9,63
11,73
9,38
6,06
-42,51
23,12 26,00 23,43 19,67
-50,22
10,78 13,92 10,05
5,77
-30,92
21,40 29,45 25,44 18,16
-22,78
51,78 52,80 50,31 49,03
-17,88
44,40 46,10 44,43 41,82
Source: CPS/FGV based on PME microdata
49
Ethnicity
Income Groups – Labor (6 Brazilian metropolitan areas)
C Class– New Middle Class (%)
Ethnicity
White
African
Asian
Mixed African-brazilian
Native indian
2008* /
2002
2002
2003
2004
Period
2005
2006
2007
2008
11,79
29,61
-1,98
26,96
6,84
45,89
39,24
35,92
41,31
49,05
46,07
38,80
35,05
39,14
47,36
47,69
39,80
38,97
41,56
44,70
50,23
44,38
40,03
46,03
48,31
51,15
46,81
36,42
49,08
48,57
50,26
47,14
35,31
50,74
49,17
51,31
50,87
35,21
52,44
52,41
2008* /
2002
2002
2003
2004
Period
2005
2006
2007
2008
18,22
30,21
13,77
28,12
29,25
16,90
27,74
15,34
24,75
29,65
15,60
23,58
15,17
22,98
22,54
E Class (%)
Ethnicity
White
African
Asian
Mixed African-brazilian
Native indian
-32,14
22,99 24,16 21,60 18,96
-38,84
38,55 39,33 36,65 32,72
-17,85
18,47 22,98 14,56 11,88
-35,26
35,50 37,91 34,99 30,95
-24,57
29,88 28,92 30,81 27,37
Source: CPS/FGV based on PME microdata
Position in the household
Income Groups – Labor (6 Brazilian metropolitan areas)
C Class– New Middle Class (%)
Position in the household
Head
Spouse
Child
Relative
Non-related member
2008* /
2002
2002
2003
2004
Period
2005
2006
2007
2008
17,63
17,46
19,24
21,56
1,21
42,45
45,30
43,80
42,74
44,96
41,97
44,31
43,35
41,29
44,54
43,60
46,44
44,99
43,78
43,81
46,89
49,88
48,41
46,25
48,94
48,40
51,36
50,76
49,08
52,16
48,31
51,37
51,15
49,70
48,53
49,94
53,22
52,23
51,95
45,51
2008* /
2002
2002
2003
2004
Period
2005
2006
2007
2008
23,58
20,04
23,43
27,27
19,30
21,47
17,72
21,43
25,76
20,20
19,71
15,94
19,97
23,30
20,66
E Class (%)
Position in the household
Head
Spouse
Child
Relative
Non-related member
-32,64
29,26 30,58 28,05 25,12
-38,00
25,71 27,45 24,40 21,40
-33,59
30,07 31,17 28,96 25,55
-29,80
33,19 36,12 33,20 29,66
-30,55
29,74 26,91 23,08 20,64
Source: CPS/IBRE/FGV based on PME microdata
50
Number of people in the household
Income Groups – Labor (6 Brazilian metropolitan areas)
C Class– New Middle Class (%)
Number of people in the
household
From 1 to 2
From 3 to 4
From 5 to 6
More than 6
2008* /
2002
2002
2003
2004
Period
2005
2006
2007
2008
14,63
20,95
23,16
20,44
45,60
42,71
37,36
20,66
45,87
41,16
34,27
17,47
47,39
43,19
35,66
18,65
50,51
46,64
38,93
19,63
51,68
49,05
41,87
23,09
50,96
50,01
43,39
24,36
52,26
51,66
46,01
24,89
2008* /
2002
2002
2003
2004
Periodo
2005
2006
2007
2008
20,03
24,63
32,97
56,50
18,52
22,12
29,69
52,79
17,23
20,09
27,36
49,06
E Class (%)
Number of people in the
household
From 1 to 2
From 3 to 4
From 5 to 6
More than 6
-31,77
25,25 26,26 23,81 21,06
-34,85
30,84 32,82 30,21 26,86
-28,53
38,28 41,03 39,00 35,99
-16,63
58,85 62,98 62,72 60,58
Source: CPS/IBRE/FGV based on PME microdata
Contributes to social security
Income Groups – Labor (6 Brazilian metropolitan areas)
C Class– New Middle Class (%)
Periodo
Contributes to social security
Yes
No
2008* /
2002
2002
2003
2004
2005
2006
2007
2008
12,31
20,19
53,52
47,71
53,72
46,44
55,58
48,78
58,78
52,11
59,88
54,83
59,36
55,12
60,11
57,34
2002
2003
2004
Periodo
2005
2006
2007
2008
7,21
19,46
6,46
17,18
5,62
15,35
E Class (%)
2008* /
2002
Contributes to social security
Yes
-54,55
12,37 13,54 10,97
8,41
No
-39,17
25,23 27,73 24,80 21,36
Source: CPS/IBRE/FGV based on PME microdata
51
APPENDIX II. REGIONAL EVALUATION
We observe that the general trend of decrease in poverty and increase of the middle
class are present in all Brazilian metropolitan areas. In general, São Paulo Metropolitan
Region has the Best rates in all years, with the lowest poverty rates, followed by the
highest proportions of individuals in the middle-class. Meanwhile, the Belo Horizonte
Metropolitan region presents the best relative performance, gaining new positions in the
rankings of (lower) poverty and middle class. Next, we detail the evolution of these two
groups in each one of the 6 metropolitan regions.
1. Poverty in the Metropolitan Regions
We present next an evolution of the poverty middle class in the main Brazilian
metropolis. As seen, when considering the last 6 years, all the cities present an
accumulated decrease of the poverty rate.
TABLE 1 - Poverty Rate
Total
Metropolitan area
Recife
Salvador
Belo Horizonte
Rio de Janeiro
São Paulo
Porto Alegre
TOTAL
2002
2003
2004
2005
2006
2007
2008*
25,20
28,74
30,17
27,60
24,49
22,78
20,70
19,00
44,06
33,88
26,48
25,73
19,93
22,91
46,64
38,19
33,32
31,22
21,90
24,49
52,26
41,97
33,56
29,38
24,00
27,81
51,89
39,05
28,51
27,16
21,70
24,25
42,99
32,59
26,03
25,10
19,30
21,97
39,15
29,99
22,37
23,59
18,38
21,25
35,91
26,27
20,09
21,29
16,83
19,93
34,64
25,12
18,64
19,74
14,78
18,36
TABLE 2 - Variation (in %) Poverty Rate
Total
Metropolitan area
Recife
Salvador
Belo Horizonte
Rio de Janeiro
São Paulo
Porto Alegre
Annual Variation
2008*/
2002
2007/
2002
2003/02
2004/03
2005/04
2006/05
2007/06
2008*/07
-33,89
-27,97
4,97
-8,51
-11,28
-6,95
-9,13
-8,22
-25,73
-34,24
-44,06
-36,75
-32,54
-25,03
-23,01
-31,22
-39,70
-31,80
-23,15
-18,64
12,05
9,89
0,71
-5,87
9,59
13,54
-0,71
-6,95
-15,05
-7,56
-9,60
-12,80
-17,14
-16,55
-8,68
-7,58
-11,07
-9,38
-8,93
-7,96
-14,07
-6,02
-4,73
-3,30
-8,29
-12,42
-10,19
-9,77
-8,44
-6,22
-3,53
-4,39
-7,22
-7,26
-12,21
-7,86
* until April
Source: CPS/FGV based on PME microdata
52
2. Middle Class
With the decreases in the rate of poverty in all metropolitan areas, the next step is to
analyze how this reflected in another population group. Therefore, we will analyze the
growth of the middle class in each one of the metropolitan areas that reach the highest
levels in all of the series now in 2008.
TABLE 3 - Proportion of the C Class – New Middle Class
Total
Metropolitan area
Recife
Salvador
Belo Horizonte
Rio de Janeiro
São Paulo
Porto Alegre
TOTAL
2002
2003
2004
2005
2006
2007
2008*
47,01
43,64
42,99
44,77
48,11
49,93
50,04
51,57
30,35
35,16
46,98
47,92
51,00
50,60
27,94
31,69
40,71
43,15
48,70
48,77
24,69
29,24
41,00
45,44
47,20
46,50
25,60
31,58
44,91
46,22
48,96
49,20
31,10
35,65
47,20
48,96
52,48
52,08
34,00
38,60
51,09
50,76
53,45
53,30
35,62
40,75
52,53
50,37
53,17
52,03
36,67
41,28
53,90
52,42
54,68
53,67
TABLE 4 - Variation (in %) in the Proportion of the C Class – New Middle Class
Total
Metropolitan area
Recife
Salvador
Belo Horizonte
Rio de Janeiro
São Paulo
Porto Alegre
Annual variation
2008*/
2002
2007/
2002
2003/02
2004/03
2005/04
2006/05
2007/06
2008*/07
18,17
14,67
-1,49
4,16
7,44
3,78
0,22
3,05
31,25
30,28
32,39
21,48
12,26
10,05
27,48
28,58
29,03
16,72
9,17
6,69
-11,63
-7,73
0,71
5,29
-3,08
-4,66
3,68
8,00
9,52
1,72
3,72
5,82
21,48
12,88
5,11
5,93
7,19
5,85
9,31
8,30
8,24
3,67
1,86
2,34
4,77
5,55
2,82
-0,77
-0,53
-2,39
2,96
1,32
2,60
4,08
2,83
3,15
* until April
Source: CPS/FGV based on PME microdata
53
APPENDIX III: Multinomial Logit Models Estimated
I) Probability of Being in Different Economic Classes
– Base (E Class)
Resposta
Class A & B
Parameter
Category
Intercept
Estimate
Standard
Error
Wald
Statistic
P-value
Odds Ratio
-6.5238
0.2336
779.6691
<.0001
.
GENDER
Male
0.6822
0.0319
456.6774
<.0001
1.9783
COR
Asian
1.6670
0.1442
133.5769
<.0001
5.2961
COR
White
1.4112
0.0661
456.0550
<.0001
4.1008
COR
Native
0.0404
0.7413
0.0030
0.9566
1.0412
COR
Afro-brazilian
0.0271
0.0706
0.1478
0.7006
1.0275
AGE
15 TO 19
-0.3849
0.0789
23.8273
<.0001
0.6805
AGE
20 TO 24
-0.5800
0.0717
65.3576
<.0001
0.5599
AGE
25 TO 29
-0.3032
0.0677
20.0362
<.0001
0.7385
AGE
30 TO 35
-0.2391
0.0650
13.5208
0.0002
0.7874
AGE
36 TO 39
-0.3057
0.0713
18.3925
<.0001
0.7366
AGE
40 TO 44
0.0264
0.0667
0.1571
0.6919
1.0268
AGE
45 TO 49
0.3506
0.0673
27.1312
<.0001
1.4199
AGE
50 TO 54
0.2865
0.0684
17.5676
<.0001
1.3318
anoest
11 OR MORE
3.6702
0.1639
501.2025
<.0001
39.2604
anoest
MISS
0.1249
0.7820
0.0255
0.8731
1.1330
anoest
1 TO 3
-0.2135
0.2088
1.0464
0.3063
0.8077
anoest
4 TO 7
0.5091
0.1699
8.9823
0.0027
1.6638
anoest
8 TO 10
1.8358
0.1670
120.8165
<.0001
6.2704
CONFAM
Housemaid
0.9877
0.2956
11.1617
0.0008
2.6849
CONFAM
Spouse
0.6985
0.0403
300.0847
<.0001
2.0107
CONFAM
Child
0.1496
0.0454
10.8802
0.0010
1.1614
CONFAM
Other relative
-0.2280
0.0847
7.2412
0.0071
0.7961
NPES
1
2.9893
0.1404
453.0755
<.0001
19.8709
NPES
2
2.2959
0.1382
275.9951
<.0001
9.9335
NPES
3
1.6377
0.1412
134.5386
<.0001
5.1436
REG
Recife
-1.3906
0.0923
227.1840
<.0001
0.2489
REG
Salvador
-0.2162
0.0787
7.5491
0.0060
0.8056
REG
Belo Horizonte
0.2517
0.0681
13.6604
0.0002
1.2862
REG
Rio de Janeiro
-0.1368
0.0572
5.7234
0.0167
0.8721
REG
São Paulo
0.4516
0.0536
70.8615
<.0001
1.5708
ANO
2003
-0.6273
0.0503
155.6717
<.0001
0.5340
ANO
2004
-0.8833
0.0496
316.6122
<.0001
0.4134
ANO
2005
-0.5676
0.0480
140.0456
<.0001
0.5669
ANO
2006
-0.4815
0.0478
101.4135
<.0001
0.6179
YEAR
2007
-0.2421
0.0472
26.2899
<.0001
0.7850
54
Class C
Intercept
-1.3926
0.0854
266.0973
<.0001
.
GENDER
Male
0.4522
0.0208
472.6205
<.0001
1.5718
COL
Asian
-0.3371
0.1296
6.7654
0.0093
0.7138
COL
White
0.2297
0.0324
50.1747
<.0001
1.2582
COL
Native
0.4194
0.2859
2.1528
0.1423
1.5211
COL
Afro-brazilian
-0.00558
0.0323
0.0299
0.8627
0.9944
AGE
15 TO 19
-0.0742
0.0496
2.2361
0.1348
0.9285
AGE
20 TO 24
0.1027
0.0471
4.7587
0.0292
1.1082
AGE
25 TO 29
0.2254
0.0454
24.6733
<.0001
1.2528
AGE
30 TO 35
0.0919
0.0430
4.5716
0.0325
1.0963
AGE
36 TO 39
0.1271
0.0462
7.5807
0.0059
1.1355
AGE
40 TO 44
0.2305
0.0441
27.2697
<.0001
1.2592
AGE
45 TO 49
0.3612
0.0452
63.8173
<.0001
1.4350
AGE
50 TO 54
0.2011
0.0457
19.4061
<.0001
1.2228
EDUCA
11 OR MORE
1.6028
0.0544
869.1895
<.0001
4.9669
EDUCA
MISS
0.4064
0.2194
3.4322
0.0639
1.5014
EDUCA
1 TO 3
0.3927
0.0621
40.0399
<.0001
1.4810
EDUCA
4 TO 7
0.6900
0.0540
163.4437
<.0001
1.9937
EDUCA
8 TO 10
1.1805
0.0553
455.9569
<.0001
3.2560
CONFAM
Housemaid
0.5939
0.2196
7.3164
0.0068
1.8111
CONFAM
Spouse
0.4703
0.0267
310.6814
<.0001
1.6004
CONFAM
Child
0.1106
0.0306
13.0621
0.0003
1.1170
CONFAM
Other relative
0.0489
0.0485
1.0149
0.3137
1.0501
NPES
1
0.7923
0.0443
320.1382
<.0001
2.2085
NPES
2
0.8848
0.0393
506.9717
<.0001
2.4225
NPES
3
0.6237
0.0412
228.6499
<.0001
1.8657
REG
Recife
-1.0869
0.0501
470.0022
<.0001
0.3373
REG
Salvador
-0.5646
0.0487
134.3251
<.0001
0.5686
REG
Belo Horizonte
-0.0228
0.0449
0.2568
0.6123
0.9775
REG
Rio de Janeiro
-0.0459
0.0389
1.3938
0.2378
0.9551
REG
São Paulo
0.1564
0.0372
17.6551
<.0001
1.1693
YEAR
2003
-0.5227
0.0338
239.3131
<.0001
0.5929
YEAR
2004
-0.7797
0.0331
553.9593
<.0001
0.4585
YEAR
2005
-0.4379
0.0328
178.0187
<.0001
0.6454
YEAR
2006
-0.3407
0.0331
106.0295
<.0001
0.7113
YEAR
2007
-0.1574
0.0333
22.3963
<.0001
0.8544
55
Class D
Intercept
-1.1023
0.1033
113.8097
<.0001
.
GENDER
Male
0.2241
0.0265
71.7247
<.0001
1.2512
COL
Asian
-1.2238
0.2581
22.4744
<.0001
0.2941
COL
White
-0.0192
0.0405
0.2261
0.6344
0.9809
COL
Native
0.1164
0.3612
0.1039
0.7472
1.1235
COL
Afro-brazilian
0.00283
0.0393
0.0052
0.9425
1.0028
AGE
15 TO 19
0.4296
0.0655
42.9797
<.0001
1.5367
AGE
20 TO 24
0.6132
0.0624
96.4759
<.0001
1.8464
AGE
25 TO 29
0.5300
0.0609
75.7098
<.0001
1.6989
AGE
30 TO 35
0.4956
0.0575
74.3061
<.0001
1.6415
AGE
36 TO 39
0.3227
0.0621
26.9741
<.0001
1.3809
AGE
40 TO 44
0.3465
0.0598
33.6003
<.0001
1.4141
AGE
45 TO 49
0.3360
0.0618
29.5260
<.0001
1.3993
AGE
50 TO 54
0.1481
0.0634
5.4608
0.0194
1.1596
EDUCA
11 OR MORE
0.3905
0.0644
36.7288
<.0001
1.4777
EDUCA
MISS
-0.1248
0.2796
0.1992
0.6554
0.8827
EDUCA
1 TO 3
0.3319
0.0721
21.1624
<.0001
1.3936
EDUCA
4 TO 7
0.3751
0.0630
35.5050
<.0001
1.4552
EDUCA
8 TO 10
0.5177
0.0649
63.6556
<.0001
1.6782
CONFAM
Housemaid
0.3041
0.2682
1.2860
0.2568
1.3555
CONFAM
Spouse
0.2310
0.0339
46.5183
<.0001
1.2598
CONFAM
Child
-0.1514
0.0392
14.9581
0.0001
0.8595
CONFAM
Other relative
-0.0852
0.0605
1.9864
0.1587
0.9183
NPES
1
-0.6166
0.0544
128.3089
<.0001
0.5398
NPES
2 es
0.1255
0.0424
8.7684
0.0031
1.1337
NPES
3 es
0.0856
0.0447
3.6579
0.0558
1.0893
REG
Recife
-0.2742
0.0607
20.3752
<.0001
0.7602
REG
Salvador
0.0214
0.0602
0.1263
0.7223
1.0216
REG
Belo Horizonte
-0.1582
0.0594
7.0945
0.0077
0.8537
REG
Rio de Janeiro
0.0538
0.0509
1.1167
0.2906
1.0553
REG
São Paulo
0.0443
0.0490
0.8160
0.3663
1.0453
YEAR
2003
-0.4655
0.0429
117.9358
<.0001
0.6279
YEAR
2004
-0.4859
0.0413
138.4398
<.0001
0.6151
YEAR
2005
-0.4122
0.0418
97.4091
<.0001
0.6622
YEAR
2006
-0.3618
0.0423
73.3239
<.0001
0.6964
YEAR
2007
-0.1372
0.0418
10.8006
0.0010
0.8718
Source: CPS/IBRE/FGV based on PME/IBGE microdata
56
b) Models with Interation dummies (YEAR*REGION)
Parameter
Category
Estimate
Standard
Error
Wald
Statistic
P-value
Odds Ratio
CLASS A & B
Intercept
-6.6631
0.2538
689.2008
<.0001
.
REG
Recife
-1.2068
0.1975
37.3217
<.0001
0.2991
REG
Salvador
-0.5404
0.1969
7.5286
0.0061
0.5825
REG
Belo Horizonte
0.3765
0.1590
5.6104
0.0179
1.4572
REG
Rio de Janeiro
0.0246
0.1312
0.0352
0.8511
1.0249
REG
São Paulo
0.6606
0.1258
27.5814
<.0001
1.9360
YEAR
2003
-0.5780
0.1708
11.4492
0.0007
0.5610
YEAR
2004
-0.5109
0.1663
9.4397
0.0021
0.5999
YEAR
2005
-0.3637
0.1717
4.4863
0.0342
0.6951
YEAR
2006
-0.3584
0.1644
4.7510
0.0293
0.6988
YEAR
2007
-0.1524
0.1616
0.8900
0.3455
0.8586
REG*YEAR
Recife
0.1311
0.3000
0.1909
0.6622
1.1400
REG*YEAR
Recife
-0.5692
0.3256
3.0548
0.0805
0.5660
REG*YEAR
Recife
-0.0673
0.2964
0.0516
0.8203
0.9349
REG*YEAR
Recife
-0.5665
0.3247
3.0441
0.0810
0.5675
REG*YEAR
Recife
-0.0124
0.2879
0.0018
0.9657
0.9877
REG*YEAR
Salvador
0.4179
0.2806
2.2183
0.1364
1.5188
REG*YEAR
Salvador
0.3026
0.2669
1.2850
0.2570
1.3533
REG*YEAR
Salvador
0.2291
0.2717
0.7111
0.3991
1.2575
REG*YEAR
Salvador
0.4658
0.2663
3.0583
0.0803
1.5932
REG*YEAR
Salvador
0.5653
0.2578
4.8069
0.0283
1.7599
REG*YEAR
Belo Horizonte
-0.1106
0.2370
0.2178
0.6407
0.8953
REG*YEAR
Belo Horizonte
-0.3999
0.2325
2.9584
0.0854
0.6704
REG*YEAR
Belo Horizonte
-0.2546
0.2312
1.2128
0.2708
0.7752
REG*YEAR
Belo Horizonte
-0.0416
0.2262
0.0338
0.8541
0.9593
REG*YEAR
Belo Horizonte
0.0669
0.2226
0.0903
0.7638
1.0692
REG*YEAR
Rio de Janeiro
-0.6103
0.2026
9.0773
0.0026
0.5432
REG*YEAR
Rio de Janeiro
-0.3173
0.1932
2.6965
0.1006
0.7281
REG*YEAR
Rio de Janeiro
-0.00413
0.1952
0.0004
0.9831
0.9959
REG*YEAR
Rio de Janeiro
-0.1446
0.1888
0.5864
0.4438
0.8654
REG*YEAR
Rio de Janeiro
-0.1226
0.1853
0.4377
0.5083
0.8846
REG*YEAR
São Paulo
0.1341
0.1864
0.5175
0.4719
1.1435
REG*YEAR
São Paulo
-0.5002
0.1823
7.5300
0.0061
0.6064
REG*YEAR
São Paulo
-0.3731
0.1864
4.0070
0.0453
0.6886
REG*YEAR
São Paulo
-0.1989
0.1798
1.2238
0.2686
0.8197
REG*YEAR
São Paulo
-0.2056
0.1772
1.3462
0.2459
0.8142
57
CLASS C
Intercept
-1.5841
0.1126
197.9410
<.0001
.
REG
Recife
-0.7739
0.1194
41.9823
<.0001
0.4612
REG
Salvador
-0.3549
0.1180
9.0506
0.0026
0.7012
REG
Belo Horizonte
0.1645
0.1112
2.1862
0.1392
1.1787
REG
Rio de Janeiro
0.0595
0.0955
0.3881
0.5333
1.0613
REG
São Paulo
0.4019
0.0934
18.4989
<.0001
1.4946
YEAR
2003
-0.3498
0.1160
9.0894
0.0026
0.7048
YEAR
2004
-0.4729
0.1166
16.4380
<.0001
0.6232
YEAR
2005
-0.0812
0.1175
0.4777
0.4895
0.9220
YEAR
2006
-0.1716
0.1165
2.1681
0.1409
0.8423
YEAR
2007
-0.0915
0.1174
0.6077
0.4356
0.9125
REG*YEAR
Recife
-0.2681
0.1675
2.5611
0.1095
0.7648
REG*YEAR
Recife
-0.9121
0.1784
26.1446
<.0001
0.4017
REG*YEAR
Recife
-0.5330
0.1671
10.1716
0.0014
0.5868
REG*YEAR
Recife
-0.3064
0.1670
3.3682
0.0665
0.7361
REG*YEAR
Recife
0.0105
0.1671
0.0039
0.9500
1.0105
REG*YEAR
Salvador
-0.1648
0.1633
1.0187
0.3128
0.8480
REG*YEAR
Salvador
-0.5476
0.1637
11.1848
0.0008
0.5783
REG*YEAR
Salvador
-0.5238
0.1631
10.3181
0.0013
0.5923
REG*YEAR
Salvador
-0.1302
0.1619
0.6468
0.4213
0.8779
REG*YEAR
Salvador
0.1018
0.1617
0.3965
0.5289
1.1072
REG*YEAR
Belo Horizonte
-0.2880
0.1538
3.5094
0.0610
0.7497
REG*YEAR
Belo Horizonte
-0.2588
0.1533
2.8504
0.0913
0.7720
REG*YEAR
Belo Horizonte
-0.4520
0.1534
8.6778
0.0032
0.6363
REG*YEAR
Belo Horizonte
-0.1540
0.1533
1.0099
0.3149
0.8572
REG*YEAR
Belo Horizonte
0.0255
0.1542
0.0273
0.8688
1.0258
REG*YEAR
Rio de Janeiro
-0.1694
0.1325
1.6358
0.2009
0.8441
REG*YEAR
Rio de Janeiro
-0.0115
0.1326
0.0076
0.9307
0.9885
REG*YEAR
Rio de Janeiro
-0.1826
0.1336
1.8678
0.1717
0.8331
REG*YEAR
Rio de Janeiro
-0.1104
0.1321
0.6983
0.4034
0.8955
REG*YEAR
Rio de Janeiro
-0.1595
0.1331
1.4351
0.2309
0.8526
REG*YEAR
São Paulo
-0.1589
0.1290
1.5165
0.2181
0.8531
REG*YEAR
São Paulo
-0.4557
0.1287
12.5323
0.0004
0.6340
REG*YEAR
São Paulo
-0.4629
0.1292
12.8382
0.0003
0.6295
REG*YEAR
São Paulo
-0.2477
0.1289
3.6894
0.0548
0.7806
REG*YEAR
São Paulo
-0.1017
0.1299
0.6128
0.4337
0.9033
58
CLASS D
Intercept
-1.3416
0.1424
88.7670
<.0001
.
REG
Recife
0.0699
0.1470
0.2260
0.6345
1.0724
REG
Salvador
0.3296
0.1450
5.1683
0.0230
1.3905
REG
Belo Horizonte
-0.0571
0.1485
0.1479
0.7006
0.9445
REG
Rio de Janeiro
0.1705
0.1258
1.8370
0.1753
1.1859
REG
São Paulo
0.3952
0.1226
10.3924
0.0013
1.4847
YEAR
2003
-0.3425
0.1571
4.7547
0.0292
0.7100
YEAR
2004
-0.1982
0.1539
1.6582
0.1978
0.8202
YEAR
2005
0.0292
0.1535
0.0361
0.8494
1.0296
YEAR
2006
-0.1580
0.1563
1.0220
0.3121
0.8538
YEAR
2007
0.1506
0.1520
0.9820
0.3217
1.1625
REG*YEAR
Recife
-0.3879
0.2090
3.4435
0.0635
0.6785
REG*YEAR
Recife
-0.4425
0.2030
4.7526
0.0293
0.6424
REG*YEAR
Recife
-0.5336
0.2018
6.9916
0.0082
0.5865
REG*YEAR
Recife
-0.2599
0.2053
1.6023
0.2056
0.7711
REG*YEAR
Recife
-0.3757
0.2039
3.3951
0.0654
0.6868
REG*YEAR
Salvador
-0.1741
0.2031
0.7344
0.3915
0.8403
REG*YEAR
Salvador
-0.7494
0.2016
13.8149
0.0002
0.4727
REG*YEAR
Salvador
-0.5124
0.1985
6.6658
0.0098
0.5990
REG*YEAR
Salvador
-0.0773
0.2002
0.1490
0.6995
0.9256
REG*YEAR
Salvador
-0.2959
0.1978
2.2388
0.1346
0.7439
REG*YEAR
Belo Horizonte
0.1284
0.2053
0.3913
0.5316
1.1370
REG*YEAR
Belo Horizonte
-0.1865
0.2033
0.8416
0.3589
0.8299
REG*YEAR
Belo Horizonte
-0.2764
0.2025
1.8637
0.1722
0.7585
REG*YEAR
Belo Horizonte
0.0170
0.2056
0.0068
0.9342
1.0171
REG*YEAR
Belo Horizonte
-0.2341
0.2036
1.3224
0.2502
0.7913
REG*YEAR
Rio de Janeiro
-0.1736
0.1783
0.9485
0.3301
0.8406
REG*YEAR
Rio de Janeiro
0.0979
0.1728
0.3208
0.5711
1.1028
REG*YEAR
Rio de Janeiro
-0.2573
0.1742
2.1818
0.1397
0.7732
REG*YEAR
Rio de Janeiro
-0.1489
0.1763
0.7138
0.3982
0.8616
REG*YEAR
Rio de Janeiro
-0.2415
0.1716
1.9798
0.1594
0.7855
REG*YEAR
São Paulo
-0.1064
0.1727
0.3798
0.5377
0.8990
REG*YEAR
São Paulo
-0.4923
0.1689
8.4953
0.0036
0.6112
REG*YEAR
São Paulo
-0.6584
0.1688
15.2221
<.0001
0.5177
REG*YEAR
São Paulo
-0.3801
0.1722
4.8696
0.0273
0.6838
REG*YEAR
São Paulo
-0.3795
0.1677
5.1238
0.0236
0.6842
Source: CPS/IBRE/FGV based on PME/IBGE microdata
59
II) Social Mobility - Probability of Entering, Exiting and Permanence Being in
Classe E – Base those that Stay in Class E
Parameter
Category
Estimate
Standar
d Error
Wald
Statistic
P-value
Odds Ratio
CONTINUE OUT OF CLASS E
Intercept
-0.7010
0.0859
66.5507
<.0001
.
GENDER
Male
0.5344
0.0220
588.6950
<.0001
1.70642
COL
Asian
0.0658
0.1325
0.2467
0.6194
1.06804
COL
White
0.2983
0.0343
75.4891
<.0001
1.34757
COL
Native
0.4955
0.3315
2.2338
0.1350
1.64135
COL
Afro-brazilian
-0.0212
0.0337
0.3951
0.5296
0.97906
AGE
15 TO 19
0.1102
0.0513
4.6119
0.0318
1.11646
AGE
20 TO 24
0.2977
0.0493
36.4707
<.0001
1.34681
AGE
25 TO 29
0.3396
0.0473
51.5281
<.0001
1.40434
AGE
30 TO 35
0.2459
0.0444
30.6513
<.0001
1.27883
AGE
36 TO 39
0.2164
0.0480
20.3271
<.0001
1.24160
AGE
40 TO 44
0.3531
0.0459
59.2935
<.0001
1.42349
AGE
45 TO 49
0.4879
0.0472
106.8374
<.0001
1.62895
AGE
50 TO 54
0.2661
0.0470
32.0624
<.0001
1.30489
EDUCA
11 OR MORE
1.8121
0.0524
1197.7525
<.0001
6.12327
EDUCA
MISS
0.1620
0.2157
0.5641
0.4526
1.17589
EDUCA
1 TO 3
0.4052
0.0597
46.1022
<.0001
1.49964
EDUCA
4 TO 7
0.6691
0.0515
168.9233
<.0001
1.95244
EDUCA
8 TO 10
1.1897
0.0533
498.7744
<.0001
3.28622
CONFAM
Housemaid
0.5707
0.2336
5.9667
0.0146
1.76950
CONFAM
Spouse
0.5379
0.0283
360.1276
<.0001
1.71244
CONFAM
Child
0.0130
0.0325
0.1586
0.6904
1.01304
CONFAM
Other relative
0.0263
0.0512
0.2648
0.6068
1.02668
NPES
1
0.6431
0.0436
217.1387
<.0001
1.90243
NPES
2 es
0.8008
0.0383
437.7412
<.0001
2.22735
NPES
3 es
0.4991
0.0403
153.2995
<.0001
1.64727
REG
Recife
-1.1197
0.0526
453.7187
<.0001
0.32638
REG
Salvador
-0.5191
0.0514
102.0210
<.0001
0.59504
REG
Belo Horizonte
-0.0761
0.0497
2.3441
0.1258
0.92673
REG
Rio de Janeiro
-0.1385
0.0431
10.3446
0.0013
0.87070
REG
São Paulo
0.1787
0.0417
18.3170
<.0001
1.19563
YEAR
2003
-0.6835
0.0371
339.7259
<.0001
0.50485
YEAR
2004
-0.8304
0.0362
527.1496
<.0001
0.43589
YEAR
2005
-0.6003
0.0359
280.1069
<.0001
0.54863
YEAR
2006
-0.5291
0.0363
212.4855
<.0001
0.58915
YEAR
2007
-0.3117
0.0365
72.8737
<.0001
0.73218
60
Parameter
Category
Estimate
Standar
d Error
Wald
Statistic
P-value
Odds Ratio
ENTERED CLASS E
Intercept
-1.9066
0.1560
149.3652
<.0001
.
GENDER
Male
0.2815
0.0394
51.0180
<.0001
1.32518
COL
Asian
-1.4179
0.3984
12.6667
0.0004
0.24223
COL
White
0.0891
0.0638
1.9484
0.1628
1.09315
COL
Native
0.8338
0.4698
3.1494
0.0760
2.30207
COL
Afro-brazilian
-0.00955
0.0635
0.0226
0.8804
0.99049
AGE
15 TO 19
0.1077
0.0950
1.2865
0.2567
1.11372
AGE
20 TO 24
0.2811
0.0906
9.6371
0.0019
1.32463
AGE
25 TO 29
0.2362
0.0877
7.2611
0.0070
1.26647
AGE
30 TO 35
0.2237
0.0826
7.3360
0.0068
1.25070
AGE
36 TO 39
0.1772
0.0890
3.9637
0.0465
1.19389
AGE
40 TO 44
0.1781
0.0861
4.2825
0.0385
1.19493
AGE
45 TO 49
0.2432
0.0885
7.5504
0.0060
1.27527
AGE
50 TO 54
0.1955
0.0884
4.8941
0.0269
1.21595
EDUCA
11 OR MORE
0.9181
0.0999
84.4287
<.0001
2.50447
EDUCA
MISS
0.2874
0.3637
0.6244
0.4294
1.33293
EDUCA
1 TO 3
0.3373
0.1135
8.8330
0.0030
1.40121
EDUCA
4 TO 7
0.5228
0.0984
28.2094
<.0001
1.68675
EDUCA
8 TO 10
0.7094
0.1016
48.7869
<.0001
2.03274
CONFAM
Housemaid
0.4650
0.3853
1.4564
0.2275
1.59202
CONFAM
Spouse
0.2968
0.0507
34.3029
<.0001
1.34558
CONFAM
Child
0.0440
0.0582
0.5713
0.4497
1.04500
CONFAM
Other relative
-0.0341
0.0952
0.1283
0.7202
0.96645
NPES
1
0.1228
0.0830
2.1910
0.1388
1.13067
NPES
2 es
0.4619
0.0706
42.7909
<.0001
1.58705
NPES
3 es
0.4360
0.0734
35.3318
<.0001
1.54657
REG
Recife
-0.5877
0.0802
53.7520
<.0001
0.55559
REG
Salvador
-1.2763
0.0940
184.1910
<.0001
0.27906
REG
Belo Horizonte
-0.3105
0.0776
16.0002
<.0001
0.73310
REG
Rio de Janeiro
-0.9977
0.0707
199.2343
<.0001
0.36872
REG
São Paulo
-0.3897
0.0642
36.8005
<.0001
0.67727
YEAR
2003
0.0717
0.0624
1.3190
0.2508
1.07435
YEAR
2004
-0.2758
0.0639
18.6346
<.0001
0.75894
YEAR
2005
-0.5016
0.0667
56.6406
<.0001
0.60556
YEAR
2006
-0.1836
0.0642
8.1722
0.0043
0.83225
YEAR
2007
-0.3170
0.0670
22.3819
<.0001
0.72835
61
EXITED CLASS E
Intercept
-1.6683
0.1428
136.4329
<.0001
.
GENDER
Male
0.3474
0.0367
89.5744
<.0001
1.41535
COL
Asian
-0.1979
0.2213
0.7992
0.3713
0.82047
COL
White
-0.0128
0.0577
0.0490
0.8248
0.98731
COL
Native
0.6975
0.4578
2.3207
0.1277
2.00865
COL
Afro-brazilian
-0.0544
0.0569
0.9124
0.3395
0.94706
AGE
15 TO 19
0.2208
0.0895
6.0843
0.0136
1.24710
AGE
20 TO 24
0.4889
0.0845
33.4422
<.0001
1.63056
AGE
25 TO 29
0.3036
0.0830
13.3885
0.0003
1.35473
AGE
30 TO 35
0.2894
0.0782
13.6869
0.0002
1.33558
AGE
36 TO 39
0.2896
0.0837
11.9766
0.0005
1.33588
AGE
40 TO 44
0.3568
0.0802
19.7860
<.0001
1.42868
AGE
45 TO 49
0.4274
0.0822
27.0397
<.0001
1.53319
AGE
50 TO 54
0.2556
0.0834
9.3810
0.0022
1.29119
EDUCA
11 OR MORE
0.8227
0.0884
86.5128
<.0001
2.27654
EDUCA
MISS
-0.2635
0.4065
0.4203
0.5168
0.76834
EDUCA
1 TO 3
0.2591
0.1008
6.6093
0.0101
1.29578
EDUCA
4 TO 7
0.3614
0.0872
17.1854
<.0001
1.43539
EDUCA
8 TO 10
0.5732
0.0901
40.4345
<.0001
1.77394
CONFAM
Housemaid
0.0184
0.4205
0.0019
0.9650
1.01860
CONFAM
Spouse
0.3249
0.0471
47.5334
<.0001
1.38395
CONFAM
Child
-0.0172
0.0541
0.1014
0.7502
0.98291
CONFAM
Other relative
0.1267
0.0835
2.3016
0.1292
1.13511
NPES
1
0.1443
0.0753
3.6720
0.0553
1.15526
NPES
2 es
0.4489
0.0645
48.4920
<.0001
1.56657
NPES
3 es
0.3014
0.0677
19.8186
<.0001
1.35179
REG
Recife
-0.6764
0.0807
70.1990
<.0001
0.50842
REG
Salvador
-0.9339
0.0861
117.5759
<.0001
0.39301
REG
Belo Horizonte
-0.2278
0.0765
8.8707
0.0029
0.79629
REG
Rio de Janeiro
-0.6927
0.0680
103.6487
<.0001
0.50021
REG
São Paulo
-0.1395
0.0633
4.8545
0.0276
0.86980
YEAR
2003
-0.2926
0.0586
24.8908
<.0001
0.74634
YEAR
2004
-0.2237
0.0561
15.8816
<.0001
0.79958
YEAR
2005
-0.5154
0.0590
76.3713
<.0001
0.59724
YEAR
2006
-0.4887
0.0599
66.5757
<.0001
0.61342
YEAR
2007
-0.5214
0.0615
71.9612
<.0001
0.59366
Source: CPS/IBRE/FGV based on PME/IBGE microdata
62
b) com dummy interativa (YEAR*REGION)
Base:Porto Alegre and 2008.
Resposta
Parameter
Category
Estimate
Standard
Error
Wald
Statistic
P-value
Odds Ratio
CONTINUE OUT OF CLASS E
Intercept
-0.9995
0.1190
70.5240
<.0001
.
REG
Recife
-0.7829
0.1297
36.4174
<.0001
0.45708
REG
Salvador
-0.2630
0.1280
4.2209
0.0399
0.76875
REG
Belo Horizonte
0.2059
0.1267
2.6405
0.1042
1.22857
REG
Rio de Janeiro
0.0507
0.1065
0.2271
0.6337
1.05205
REG
São Paulo
0.6053
0.1066
32.2165
<.0001
1.83180
YEAR
2003
-0.3854
0.1321
8.5095
0.0035
0.68015
YEAR
2004
-0.3202
0.1332
5.7762
0.0162
0.72597
YEAR
2005
-0.2227
0.1305
2.9132
0.0879
0.80035
YEAR
2006
-0.2740
0.1297
4.4633
0.0346
0.76030
YEAR
2007
-0.0712
0.1310
0.2951
0.5869
0.93129
REG*YEAR
Recife
-0.4045
0.1801
5.0429
0.0247
0.66731
REG*YEAR
Recife
-0.8432
0.1825
21.3438
<.0001
0.43031
REG*YEAR
Recife
-0.3679
0.1753
4.4055
0.0358
0.69217
REG*YEAR
Recife
-0.3149
0.1775
3.1498
0.0759
0.72983
REG*YEAR
Recife
-0.0169
0.1804
0.0088
0.9253
0.98323
REG*YEAR
Salvador
-0.2531
0.1752
2.0859
0.1487
0.77641
REG*YEAR
Salvador
-0.7809
0.1744
20.0447
<.0001
0.45798
REG*YEAR
Salvador
-0.3465
0.1725
4.0333
0.0446
0.70714
REG*YEAR
Salvador
-0.0248
0.1730
0.0206
0.8859
0.97549
REG*YEAR
Salvador
-0.0946
0.1729
0.2998
0.5840
0.90970
REG*YEAR
Belo Horizonte
-0.4851
0.1728
7.8811
0.0050
0.61561
REG*YEAR
Belo Horizonte
-0.3894
0.1742
4.9968
0.0254
0.67745
REG*YEAR
Belo Horizonte
-0.4741
0.1703
7.7535
0.0054
0.62245
REG*YEAR
Belo Horizonte
-0.1885
0.1713
1.2113
0.2711
0.82817
REG*YEAR
Belo Horizonte
-0.0997
0.1736
0.3297
0.5658
0.90513
REG*YEAR
Rio de Janeiro
-0.2546
0.1491
2.9151
0.0878
0.77520
REG*YEAR
Rio de Janeiro
-0.1935
0.1493
1.6790
0.1951
0.82410
REG*YEAR
Rio de Janeiro
-0.1922
0.1468
1.7152
0.1903
0.82511
REG*YEAR
Rio de Janeiro
-0.1925
0.1456
1.7488
0.1860
0.82488
REG*YEAR
Rio de Janeiro
-0.3040
0.1467
4.2910
0.0383
0.73788
REG*YEAR
São Paulo
-0.3298
0.1475
4.9992
0.0254
0.71909
REG*YEAR
São Paulo
-0.7377
0.1475
25.0249
<.0001
0.47820
REG*YEAR
São Paulo
-0.5632
0.1448
15.1372
<.0001
0.56940
REG*YEAR
São Paulo
-0.4364
0.1447
9.0909
0.0026
0.64635
REG*YEAR
São Paulo
-0.3864
0.1462
6.9876
0.0082
0.67951
63
Resposta
Parameter
Category
Estimate
Standard
Error
Wald
Statistic
P-value
Odds Ratio
ENTERED CLASS E
Intercept
REG
Recife
REG
Salvador
REG
-2.3248
0.2073
125.7219
<.0001
.
0.2386
0.1973
1.4635
0.2264
1.26953
-0.4249
0.2233
3.6211
0.0571
0.65382
Belo Horizonte
0.1667
0.2045
0.6645
0.4150
1.18143
REG
Rio de Janeiro
-1.1758
0.2024
33.7388
<.0001
0.30857
REG
São Paulo
0.1234
0.1751
0.4970
0.4808
1.13136
YEAR
2003
0.4596
0.1990
5.3331
0.0209
1.58341
YEAR
2004
0.2567
0.2066
1.5432
0.2141
1.29267
YEAR
2005
0.2970
0.2023
2.1568
0.1419
1.34587
YEAR
2006
0.1428
0.2052
0.4842
0.4865
1.15350
YEAR
2007
0.0114
0.2137
0.0029
0.9574
1.01148
REG*YEAR
Recife
-0.9960
0.2610
14.5630
0.0001
0.36937
REG*YEAR
Recife
-1.2494
0.2733
20.8935
<.0001
0.28669
REG*YEAR
Recife
-1.6819
0.2830
35.3322
<.0001
0.18602
REG*YEAR
Recife
-0.6978
0.2672
6.8205
0.0090
0.49766
REG*YEAR
Recife
-0.3321
0.2776
1.4308
0.2316
0.71742
REG*YEAR
Salvador
-0.7866
0.2941
7.1553
0.0075
0.45540
REG*YEAR
Salvador
-1.1613
0.3080
14.2189
0.0002
0.31307
REG*YEAR
Salvador
-1.7066
0.3461
24.3114
<.0001
0.18149
REG*YEAR
Salvador
-0.6049
0.3068
3.8878
0.0486
0.54613
REG*YEAR
Salvador
-1.0365
0.3368
9.4719
0.0021
0.35470
REG*YEAR
Belo Horizonte
-0.4688
0.2606
3.2349
0.0721
0.62577
REG*YEAR
Belo Horizonte
-0.8110
0.2785
8.4802
0.0036
0.44440
REG*YEAR
Belo Horizonte
-0.8707
0.2720
10.2466
0.0014
0.41866
REG*YEAR
Belo Horizonte
-0.4298
0.2731
2.4774
0.1155
0.65061
REG*YEAR
Belo Horizonte
-0.2581
0.2829
0.8325
0.3616
0.77250
REG*YEAR
Rio de Janeiro
0.2888
0.2531
1.3024
0.2538
1.33485
REG*YEAR
Rio de Janeiro
0.5425
0.2576
4.4355
0.0352
1.72024
REG*YEAR
Rio de Janeiro
-0.1585
0.2635
0.3617
0.5476
0.85346
REG*YEAR
Rio de Janeiro
0.0274
0.2633
0.0108
0.9172
1.02777
REG*YEAR
Rio de Janeiro
0.0844
0.2719
0.0965
0.7561
1.08810
REG*YEAR
São Paulo
-0.4780
0.2253
4.5034
0.0338
0.62000
REG*YEAR
São Paulo
-0.7610
0.2329
10.6779
0.0011
0.46721
REG*YEAR
São Paulo
-0.9042
0.2301
15.4412
<.0001
0.40488
REG*YEAR
São Paulo
-0.3606
0.2312
2.4334
0.1188
0.69723
REG*YEAR
São Paulo
-0.4440
0.2414
3.3828
0.0659
0.64144
64
Resposta
Parameter
Category
Estimate
Standard
Error
Wald
Statistic
P-value
Odds Ratio
EXITED CLASS E
Intercept
-2.0984
0.1938
117.2159
<.0001
.
REG
Recife
-0.2001
0.2008
0.9933
0.3189
0.81867
REG
Salvador
-0.3185
0.2077
2.3505
0.1252
0.72724
REG
Belo Horizonte
0.2356
0.1941
1.4729
0.2249
1.26562
REG
Rio de Janeiro
-0.3200
0.1719
3.4654
0.0627
0.72614
REG
São Paulo
0.3791
0.1652
5.2681
0.0217
1.46102
YEAR
2003
0.2971
0.1956
2.3060
0.1289
1.34595
YEAR
2004
0.4467
0.1942
5.2908
0.0214
1.56319
YEAR
2005
-0.2054
0.2104
0.9525
0.3291
0.81436
YEAR
2006
-0.1232
0.2058
0.3584
0.5494
0.88409
YEAR
2007
-0.0464
0.2074
0.0501
0.8228
0.95463
REG*YEAR
Recife
-0.6803
0.2676
6.4650
0.0110
0.50647
REG*YEAR
Recife
-0.9778
0.2661
13.4997
0.0002
0.37613
REG*YEAR
Recife
-0.6321
0.2863
4.8740
0.0273
0.53148
REG*YEAR
Recife
-0.3777
0.2794
1.8278
0.1764
0.68544
REG*YEAR
Recife
0.0268
0.2788
0.0093
0.9234
1.02718
REG*YEAR
Salvador
-0.9573
0.2859
11.2082
0.0008
0.38393
REG*YEAR
Salvador
-1.1942
0.2770
18.5856
<.0001
0.30295
REG*YEAR
Salvador
-0.2484
0.2881
0.7432
0.3886
0.78008
REG*YEAR
Salvador
-0.1767
0.2857
0.3826
0.5362
0.83801
REG*YEAR
Salvador
-1.1262
0.3185
12.4999
0.0004
0.32426
REG*YEAR
Belo Horizonte
-0.9991
0.2626
14.4727
0.0001
0.36819
REG*YEAR
Belo Horizonte
-0.5308
0.2528
4.4100
0.0357
0.58812
REG*YEAR
Belo Horizonte
-0.5028
0.2729
3.3960
0.0654
0.60482
REG*YEAR
Belo Horizonte
-0.3728
0.2696
1.9122
0.1667
0.68884
REG*YEAR
Belo Horizonte
-0.2692
0.2713
0.9848
0.3210
0.76396
REG*YEAR
Rio de Janeiro
-0.2282
0.2273
1.0084
0.3153
0.79594
REG*YEAR
Rio de Janeiro
-0.5432
0.2266
5.7448
0.0165
0.58091
REG*YEAR
Rio de Janeiro
-0.3766
0.2468
2.3296
0.1269
0.68617
REG*YEAR
Rio de Janeiro
-0.4454
0.2410
3.4163
0.0646
0.64056
REG*YEAR
Rio de Janeiro
-0.6373
0.2445
6.7926
0.0092
0.52870
REG*YEAR
São Paulo
-0.7505
0.2193
11.7097
0.0006
0.47215
REG*YEAR
São Paulo
-0.7942
0.2151
13.6361
0.0002
0.45193
REG*YEAR
São Paulo
-0.3028
0.2307
1.7220
0.1894
0.73876
REG*YEAR
São Paulo
-0.4879
0.2285
4.5599
0.0327
0.61392
REG*YEAR
São Paulo
-0.6135
0.2312
7.0416
0.0080
0.54146
Source: CPS/IBRE/FGV based on PME/IBGE microdata
65
III) Mobility (Entry and Exit of Class E ) Resposta
Parameter
Category
Estimate
Standar
d Error
Wald
Statistic
P-value
Odds Ratio
CONTINUE OUT OF CLASS E
Intercept
1.5723
0.0814
372.6682
<.0001
.
GENDER
Male
-0.2583
0.0181
203.4517
<.0001
0.77237
COL
Asian
1.0103
0.1050
92.6249
<.0001
2.74652
COL
White
0.0989
0.0291
11.5502
0.0007
1.10391
COL
Native
-0.1902
0.2757
0.4757
0.4904
0.82682
COL
Afro-brazilian
0.0242
0.0294
0.6779
0.4103
1.02453
AGE
15 TO 19
0.1827
0.0444
16.9177
<.0001
1.20051
AGE
20 TO 24
-0.1345
0.0415
10.4887
0.0012
0.87413
AGE
25 TO 29
-0.1591
0.0396
16.1146
<.0001
0.85290
AGE
30 TO 35
-0.00013
0.0381
0.0000
0.9973
0.99987
AGE
36 TO 39
-0.1053
0.0410
6.5973
0.0102
0.90005
AGE
40 TO 44
-0.1293
0.0390
10.9621
0.0009
0.87875
AGE
45 TO 49
-0.2043
0.0398
26.3844
<.0001
0.81519
AGE
50 TO 54
-0.1172
0.0407
8.2949
0.0040
0.88943
EDUCA
11 OR MORE
-0.9111
0.0545
279.5938
<.0001
0.40207
EDUCA
MISS
-0.2422
0.2306
1.1033
0.2935
0.78489
EDUCA
1 TO 3
-0.4118
0.0629
42.8457
<.0001
0.66248
EDUCA
4 TO 7
-0.6483
0.0550
138.7110
<.0001
0.52296
EDUCA
8 TO 10
-1.0154
0.0557
331.8562
<.0001
0.36227
CONFAM
Housemaid
-0.1677
0.1697
0.9764
0.3231
0.84564
CONFAM
Spouse
-0.2743
0.0230
142.7288
<.0001
0.76012
CONFAM
Child
-0.1935
0.0263
54.0140
<.0001
0.82404
CONFAM
Other relative
-0.1484
0.0436
11.5998
0.0007
0.86211
NPES
1
-0.6381
0.0420
230.3326
<.0001
0.52831
NPES
2 es
-0.8462
0.0384
486.8656
<.0001
0.42904
NPES
3 es
-0.6830
0.0401
290.5224
<.0001
0.50509
REG
Recife
1.1455
0.0488
551.7870
<.0001
3.14389
REG
Salvador
0.6525
0.0435
225.4463
<.0001
1.92041
REG
Belo Horizonte
0.1207
0.0400
9.1224
0.0025
1.12826
REG
Rio de Janeiro
0.0779
0.0338
5.3065
0.0212
1.08104
REG
São Paulo
0.000543
0.0324
0.0003
0.9866
1.00054
YEAR
2003
0.3265
0.0290
126.8949
<.0001
1.38604
YEAR
2004
0.4884
0.0286
291.5099
<.0001
1.62973
YEAR
2005
0.2124
0.0273
60.5157
<.0001
1.23665
YEAR
2006
0.1557
0.0274
32.2816
<.0001
1.16847
YEAR
2007
0.0862
0.0270
10.2309
0.0014
1.09003
66
ENTERED CLASS E
Intercept
-0.8878
0.1471
36.3984
<.0001
.
GENDER
Male
-0.0852
0.0316
7.2838
0.0070
0.91833
COL
Asian
0.5765
0.1746
10.9008
0.0010
1.77986
COL
White
0.0120
0.0524
0.0527
0.8184
1.01211
COL
Native
0.1575
0.4369
0.1300
0.7185
1.17059
COL
Afro-brazilian
0.0295
0.0535
0.3037
0.5816
1.02993
AGE
15 TO 19
0.1363
0.0798
2.9185
0.0876
1.14607
AGE
20 TO 24
0.1856
0.0736
6.3645
0.0116
1.20397
AGE
25 TO 29
-0.0508
0.0722
0.4953
0.4816
0.95047
AGE
30 TO 35
0.0331
0.0696
0.2259
0.6346
1.03361
AGE
36 TO 39
-0.0118
0.0748
0.0248
0.8748
0.98829
AGE
40 TO 44
0.0355
0.0707
0.2521
0.6156
1.03616
AGE
45 TO 49
0.0816
0.0711
1.3159
0.2513
1.08501
AGE
50 TO 54
0.0746
0.0734
1.0325
0.3096
1.07746
EDUCA
11 OR MORE
-0.3593
0.0991
13.1527
0.0003
0.69818
EDUCA
MISS
-0.8260
0.5494
2.2602
0.1327
0.43779
EDUCA
1 TO 3
0.0500
0.1120
0.1993
0.6553
1.05129
EDUCA
4 TO 7
-0.2040
0.1000
4.1600
0.0414
0.81549
EDUCA
8 TO 10
-0.3652
0.1011
13.0598
0.0003
0.69407
CONFAM
Housemaid
-0.8523
0.4249
4.0236
0.0449
0.42642
CONFAM
Spouse
-0.1033
0.0404
6.5232
0.0106
0.90186
CONFAM
Child
0.0167
0.0458
0.1326
0.7158
1.01681
CONFAM
Other relative
0.0500
0.0753
0.4417
0.5063
1.05129
NPES
1
0.0434
0.0817
0.2823
0.5952
1.04438
NPES
2 es
0.0673
0.0752
0.8006
0.3709
1.06964
NPES
3 es
-0.0837
0.0786
1.1342
0.2869
0.91971
REG
Recife
0.3507
0.0758
21.3877
<.0001
1.42003
REG
Salvador
-0.6334
0.0800
62.7296
<.0001
0.53079
REG
Belo Horizonte
-0.1762
0.0618
8.1350
0.0043
0.83842
REG
Rio de Janeiro
-0.7392
0.0543
185.5055
<.0001
0.47750
REG
São Paulo
-0.3480
0.0490
50.5037
<.0001
0.70608
YEAR
2003
0.0814
0.0476
2.9197
0.0875
1.08481
YEAR
2004
0.2673
0.0461
33.6806
<.0001
1.30645
YEAR
2005
-0.3292
0.0483
46.4616
<.0001
0.71949
YEAR
2006
-0.2388
0.0472
25.5815
<.0001
0.78754
YEAR
2007
-0.3270
0.0469
48.6995
<.0001
0.72110
67
EXITED CLASS C
Intercept
-0.8390
0.1475
32.3761
<.0001
.
GENDER
Male
-0.1110
0.0323
11.7873
0.0006
0.89492
COL
Asian
0.6437
0.1813
12.6050
0.0004
1.90344
COL
White
0.1241
0.0555
5.0018
0.0253
1.13216
COL
Native
0.9166
0.3453
7.0448
0.0079
2.50084
COL
Afro-brazilian
0.1382
0.0563
6.0255
0.0141
1.14825
AGE
15 TO 19
0.0184
0.0810
0.0516
0.8203
1.01858
AGE
20 TO 24
-0.0628
0.0757
0.6884
0.4067
0.93909
AGE
25 TO 29
-0.0493
0.0725
0.4632
0.4961
0.95188
AGE
30 TO 35
-0.0159
0.0703
0.0510
0.8213
0.98425
AGE
36 TO 39
-0.0126
0.0750
0.0284
0.8663
0.98745
AGE
40 TO 44
-0.0217
0.0716
0.0920
0.7617
0.97851
AGE
45 TO 49
-0.0249
0.0727
0.1174
0.7319
0.97541
AGE
50 TO 54
-0.0493
0.0755
0.4274
0.5133
0.95185
EDUCA
11 OR MORE
-0.2909
0.1016
8.1985
0.0042
0.74762
EDUCA
MISS
0.3476
0.3568
0.9490
0.3300
1.41560
EDUCA
1 TO 3
-0.1082
0.1172
0.8525
0.3558
0.89746
EDUCA
4 TO 7
-0.1562
0.1025
2.3247
0.1273
0.85535
EDUCA
8 TO 10
-0.3339
0.1037
10.3748
0.0013
0.71610
CONFAM
Housemaid
-0.0983
0.3118
0.0994
0.7525
0.90634
CONFAM
Spouse
-0.1159
0.0414
7.8520
0.0051
0.89055
CONFAM
Child
0.0237
0.0468
0.2577
0.6117
1.02403
CONFAM
Other relative
0.0257
0.0779
0.1089
0.7413
1.02603
NPES
1
-0.3780
0.0766
24.3754
<.0001
0.68522
NPES
2 es
-0.2889
0.0686
17.7505
<.0001
0.74910
NPES
3 es
-0.3618
0.0721
25.2116
<.0001
0.69639
REG
Recife
0.6395
0.0744
73.8645
<.0001
1.89549
REG
Salvador
-0.4604
0.0803
32.8630
<.0001
0.63105
REG
Belo Horizonte
0.0287
0.0624
0.2112
0.6458
1.02908
REG
Rio de Janeiro
-0.7085
0.0568
155.8132
<.0001
0.49238
REG
São Paulo
-0.3532
0.0513
47.4403
<.0001
0.70242
YEAR
2003
0.3656
0.0483
57.3732
<.0001
1.44136
YEAR
2004
0.2441
0.0498
23.9783
<.0001
1.27642
YEAR
2005
-0.2132
0.0507
17.6668
<.0001
0.80796
YEAR
2006
-0.0428
0.0488
0.7687
0.3806
0.95814
YEAR
2007
-0.1488
0.0486
9.3742
0.0022
0.86177
Source: CPS/IBRE/FGV based on PME/IBGE microdata
68
b) Mobility (Entry and Exit of Class E )
Models with Interation dummies (YEAR*REGION)
Resposta
Parameter
Category
Estimate
Standard
Error
Wald
Statistic
P-value
Odds Ratio
CONTINUE OUT OF CLASS C
Intercept
1.6710
0.1017
269.9544
<.0001
.
REG
Recife
0.8301
0.1096
57.3526
<.0001
2.29348
REG
Salvador
0.4726
0.1011
21.8632
<.0001
1.60416
REG
Belo Horizonte
-0.1076
0.0951
1.2798
0.2579
0.89803
REG
Rio de Janeiro
0.0301
0.0785
0.1469
0.7015
1.03053
REG
São Paulo
-0.0532
0.0764
0.4848
0.4863
0.94819
YEAR
2003
0.1538
0.1034
2.2090
0.1372
1.16621
YEAR
2004
0.2924
0.1037
7.9545
0.0048
1.33963
YEAR
2005
0.0618
0.0998
0.3833
0.5359
1.06375
YEAR
2006
0.0776
0.0990
0.6138
0.4333
1.08067
YEAR
2007
0.1148
0.0989
1.3468
0.2458
1.12165
REG*YEAR
Recife
0.4992
0.1664
9.0000
0.0027
1.64737
REG*YEAR
Recife
0.8879
0.1775
25.0213
<.0001
2.43006
REG*YEAR
Recife
0.3682
0.1554
5.6173
0.0178
1.44515
REG*YEAR
Recife
0.4018
0.1599
6.3156
0.0120
1.49449
REG*YEAR
Recife
-0.0601
0.1552
0.1502
0.6984
0.94163
REG*YEAR
Salvador
0.3259
0.1482
4.8379
0.0278
1.38531
REG*YEAR
Salvador
0.4491
0.1492
9.0626
0.0026
1.56695
REG*YEAR
Salvador
0.3370
0.1419
5.6365
0.0176
1.40075
REG*YEAR
Salvador
0.1578
0.1405
1.2603
0.2616
1.17089
REG*YEAR
Salvador
-0.0895
0.1375
0.4242
0.5148
0.91436
REG*YEAR
Belo Horizonte
0.5057
0.1389
13.2653
0.0003
1.65817
REG*YEAR
Belo Horizonte
0.2653
0.1379
3.7017
0.0544
1.30378
REG*YEAR
Belo Horizonte
0.3910
0.1334
8.5909
0.0034
1.47842
REG*YEAR
Belo Horizonte
0.2448
0.1326
3.4051
0.0650
1.27731
REG*YEAR
Belo Horizonte
-0.0222
0.1323
0.0282
0.8667
0.97802
REG*YEAR
Rio de Janeiro
0.0178
0.1171
0.0231
0.8792
1.01796
REG*YEAR
Rio de Janeiro
0.0316
0.1161
0.0739
0.7857
1.03208
REG*YEAR
Rio de Janeiro
0.1184
0.1121
1.1157
0.2908
1.12569
REG*YEAR
Rio de Janeiro
0.0280
0.1110
0.0634
0.8012
1.02835
REG*YEAR
Rio de Janeiro
0.0877
0.1106
0.6285
0.4279
1.09167
REG*YEAR
São Paulo
0.1467
0.1127
1.6956
0.1929
1.15800
REG*YEAR
São Paulo
0.1995
0.1129
3.1215
0.0773
1.22076
REG*YEAR
São Paulo
0.0712
0.1084
0.4316
0.5112
1.07381
REG*YEAR
São Paulo
0.0369
0.1080
0.1172
0.7321
1.03764
REG*YEAR
São Paulo
-0.0973
0.1075
0.8188
0.3655
0.90727
69
Resposta
Parameter
Category
Estimate
Standard
Error
Wald
Statistic
P-value
Odds Ratio
ENTERED CLASS C
Intercept
-1.0433
0.1717
36.9384
<.0001
.
0.4193
0.1620
6.6986
0.0096
1.52097
-0.2966
0.1677
3.1292
0.0769
0.74333
REG
Recife
REG
Salvador
REG
Belo Horizonte
0.0512
0.1361
0.1419
0.7064
1.05258
REG
Rio de Janeiro
-0.8450
0.1259
45.0251
<.0001
0.42956
REG
São Paulo
-0.0859
0.1121
0.5868
0.4437
0.91773
YEAR
2003
0.2391
0.1477
2.6207
0.1055
1.27009
YEAR
2004
0.4217
0.1451
8.4409
0.0037
1.52450
YEAR
2005
-0.2645
0.1558
2.8824
0.0896
0.76758
YEAR
2006
-0.2035
0.1529
1.7697
0.1834
0.81591
YEAR
2007
0.1294
0.1434
0.8136
0.3671
1.13810
REG*YEAR
Recife
-0.1844
0.2499
0.5446
0.4605
0.83158
REG*YEAR
Recife
0.2514
0.2496
1.0147
0.3138
1.28586
REG*YEAR
Recife
-0.3125
0.2592
1.4536
0.2280
0.73163
REG*YEAR
Recife
0.2416
0.2459
0.9653
0.3259
1.27326
REG*YEAR
Recife
-0.2508
0.2321
1.1671
0.2800
0.77820
REG*YEAR
Salvador
-0.4396
0.2602
2.8538
0.0912
0.64430
REG*YEAR
Salvador
-0.2387
0.2485
0.9230
0.3367
0.78766
REG*YEAR
Salvador
-0.1562
0.2627
0.3537
0.5520
0.85537
REG*YEAR
Salvador
0.0900
0.2442
0.1357
0.7126
1.09413
REG*YEAR
Salvador
-1.4263
0.2826
25.4756
<.0001
0.24020
REG*YEAR
Belo Horizonte
-0.3599
0.2066
3.0364
0.0814
0.69771
REG*YEAR
Belo Horizonte
-0.5290
0.2009
6.9316
0.0085
0.58921
REG*YEAR
Belo Horizonte
0.0135
0.2074
0.0042
0.9481
1.01361
REG*YEAR
Belo Horizonte
-0.1256
0.2050
0.3758
0.5399
0.88193
REG*YEAR
Belo Horizonte
-0.3696
0.1930
3.6699
0.0554
0.69098
REG*YEAR
Rio de Janeiro
0.4228
0.1784
5.6144
0.0178
1.52627
REG*YEAR
Rio de Janeiro
0.2433
0.1751
1.9305
0.1647
1.27545
REG*YEAR
Rio de Janeiro
0.1377
0.1918
0.5152
0.4729
1.14759
REG*YEAR
Rio de Janeiro
0.1597
0.1863
0.7344
0.3915
1.17315
REG*YEAR
Rio de Janeiro
-0.3962
0.1823
4.7256
0.0297
0.67286
REG*YEAR
São Paulo
-0.3771
0.1640
5.2862
0.0215
0.68583
REG*YEAR
São Paulo
-0.2975
0.1607
3.4288
0.0641
0.74268
REG*YEAR
São Paulo
-0.1467
0.1706
0.7386
0.3901
0.86359
REG*YEAR
São Paulo
-0.1444
0.1681
0.7372
0.3905
0.86557
REG*YEAR
São Paulo
-0.5782
0.1592
13.1975
0.0003
0.56091
70
Resposta
Parameter
Category
Estimate
Standard
Error
Wald
Statistic
P-value
Odds Ratio
EXITED CLASS C
Intercept
-1.0248
0.1781
33.1248
<.0001
.
0.8246
0.1647
25.0649
<.0001
2.28093
-0.0392
0.1754
0.0501
0.8230
0.96152
REG
Recife
REG
Salvador
REG
Belo Horizonte
0.2606
0.1459
3.1932
0.0739
1.29774
REG
Rio de Janeiro
-1.0888
0.1468
55.0004
<.0001
0.33664
REG
São Paulo
-0.0127
0.1240
0.0105
0.9184
0.98738
YEAR
2003
0.4742
0.1564
9.1937
0.0024
1.60665
YEAR
2004
0.4081
0.1610
6.4276
0.0112
1.50403
YEAR
2005
0.1333
0.1592
0.7016
0.4023
1.14262
YEAR
2006
0.0164
0.1622
0.0103
0.9193
1.01657
YEAR
2007
0.2060
0.1567
1.7281
0.1886
1.22880
REG*YEAR
Recife
-0.0761
0.2381
0.1023
0.7491
0.92669
REG*YEAR
Recife
-0.5862
0.2824
4.3087
0.0379
0.55642
REG*YEAR
Recife
-0.7193
0.2521
8.1427
0.0043
0.48710
REG*YEAR
Recife
0.2345
0.2389
0.9636
0.3263
1.26427
REG*YEAR
Recife
-0.1848
0.2299
0.6465
0.4214
0.83126
REG*YEAR
Salvador
-0.1194
0.2451
0.2373
0.6262
0.88745
REG*YEAR
Salvador
-0.4604
0.2680
2.9507
0.0858
0.63103
REG*YEAR
Salvador
-1.0792
0.2962
13.2711
0.0003
0.33987
REG*YEAR
Salvador
-0.2958
0.2567
1.3276
0.2492
0.74396
REG*YEAR
Salvador
-0.8147
0.2541
10.2812
0.0013
0.44277
REG*YEAR
Belo Horizonte
-0.2420
0.2070
1.3670
0.2423
0.78504
REG*YEAR
Belo Horizonte
-0.5231
0.2143
5.9572
0.0147
0.59267
REG*YEAR
Belo Horizonte
-0.2816
0.2095
1.8078
0.1788
0.75454
REG*YEAR
Belo Horizonte
-0.1121
0.2090
0.2878
0.5917
0.89392
REG*YEAR
Belo Horizonte
-0.2678
0.2016
1.7651
0.1840
0.76502
REG*YEAR
Rio de Janeiro
0.6575
0.1947
11.4106
0.0007
1.93004
REG*YEAR
Rio de Janeiro
0.6205
0.1982
9.7987
0.0017
1.85993
REG*YEAR
Rio de Janeiro
0.4751
0.1994
5.6745
0.0172
1.60816
REG*YEAR
Rio de Janeiro
0.4275
0.2026
4.4546
0.0348
1.53343
REG*YEAR
Rio de Janeiro
-0.0846
0.2034
0.1730
0.6774
0.91888
REG*YEAR
São Paulo
-0.3736
0.1725
4.6919
0.0303
0.68823
REG*YEAR
São Paulo
-0.2915
0.1773
2.7016
0.1002
0.74714
REG*YEAR
São Paulo
-0.6298
0.1768
12.6890
0.0004
0.53267
REG*YEAR
São Paulo
-0.2232
0.1779
1.5744
0.2096
0.79996
REG*YEAR
São Paulo
-0.5489
0.1728
10.0867
0.0015
0.57759
Source: CPS/IBRE/FGV based on PME/IBGE microdata
71