An analysis of the earnings in Brazilian metropolitan areas by color - - PDF document

SLIDE 1

1

An analysis of the earnings in Brazilian metropolitan areas by color and gender from the view of mobility in the 21 st century1 Mariangela Furlan Antigo – Universidade Federal de Minas Gerais Daniela Goes Paraiso Lacerda – Fundação João Pinheiro Carolina Guinesi Mattos Borges– Universidade Federal de Minas Gerais Sandro Eduardo Monsueto – Universidade Federal de Goiás Abstract Labor income inequality in Brazil dropped in the first decade of the XXI century, benefiting the population of six Brazilian metropolitan regions contemplated in the analysis, namely, Belo Horizonte, Porto Alegre, Recife, Rio de Janeiro, Salvador e São

• Paulo. Despite this scenario of improvement, differences are still seen when individuals

are considered by gender and color. Based on possible relations between distribution and income mobility, an analysis is carried out on how mobility, at different points of distribution and combined with gender and color traits, have an influence on the reduction or maintenance of inequality. By using data from the Monthly Employment Survey, the results of mobility indicators and quantile regressions have shown that mobility contributed for a more deconcentrated distribution throughout time, in different levels for white and non-white. Non-white have registered higher mobility, especially women, along with smaller levels of inequality and higher concentration at the basis of the distribution. Furthermore, higher levels of ascending mobility for non-white at the basis of distribution might contribute for lesser levels of inequality between groups in the long term. Keywords: mobility; inequality; income; gender; color Introduction The analysis of income mobility allows quantifying how the position of individuals in the distribution alters throughout time, recognizing that the present position is dependent of its previous configuration in the past period. These movements reflect wellbeing of individuals, once higher positions in society tend to be followed by enhancement of this factor. A more deconcentrated income distribution is possible by two means, either by gains for those situated at the basis of the pyramid (ascending income mobility) or by losses for those situated at the top (descending mobility). Analyzing the Brazilian case, it is noticeable that work income inequality in Brazil dropped in the first decade of the XXI century, benefiting the population of six Brazilian metropolitan regions contemplated in the analysis, namely, Belo Horizonte, Porto Alegre, Recife, Rio de Janeiro, Salvador e São Paulo. Despite this scenario of improvement, differences are still seen when individuals are considered by gender and

1 This study had financial support of Fapemig (Edital 01/2014 – Demanda Universal, Process APQ-

02764-14 and Edital 02/2016 - Programa Pesquisador Mineiro - PPM X, Process PPM-00658-16).

SLIDE 2

2

• color. If groups present different patterns of mobility, albeit keep the same level of

inequality, it is possible that those with higher mobility might present some alterations regarding the individuals located at the top and at the basis of the pyramid. Ascending mobility in the basis, for example, might compensate inequality throughout time, fostering income dislocation amongst those with lower income level. Given that, based on the possible relations between distribution and income mobility and non-homogeneity of Brazilian sub-groups regarding income, it is desirable to evaluate how individual factors, namely, gender and color, might explain part of the mobility between different distribution strata and its relation with inequality. In general terms, mobility indicators and quantile regressions, obtained from Monthly Employment Survey (PME) from the Brazilian Institute of Geography and Statistics (IBGE), shown that mobility contributes for a more deconcentrated distribution throughout time, in different levels for white and non-white. The non-white has registered more mobility, mostly for women, along with lower levels of inequality and higher concentration at the basis of the wage pyramid. Further, higher levels of ascending mobility for non-white at the basis of distribution might contribute for lower inequality between groups in the long term. Theoretical Framework The income distribution portrays the concentration of individuals in different levels of income in each time. Its usual measurement in a point in time or the use of repetitive cross section does not allow inferences about the dynamic of income patterns. To visualize the difference between the inequality and mobility analysis, consider two societies A and B, given that A shows a more stiff structure than B. Suppose that the cities present different patterns of mobility, yet keep the same indicators of income inequality in time. In society A, poorer individuals will continue to be poor and rich will continue in the same point of distribution, meanwhile, society B may present change in the composition of poor and rich that is not depicted by inequality

• measurements. In case the same individual in society was considered throughout time, it

would be possible to verify, for example, if those initially poorer stays on that position

• r if, for instance, the economic growth would originally benefit more the poorer

individuals than the richer. The length in which high inequality can be seen as a minor consequence for society has been recently a major object of study. That occurs as long as it is accompanied by increasing change of intergenerational mobility (CORAK, 2004). If that happens, the mobility compensates inequality as far as income variations furthers income movement for those positioned in the basis of distribution. Gottschalk e Moffitt (1994), Gittleman e Joyce (1996) e Buchinsky e Hunt (1999) research if the increase of income inequality has been accompanied by the increase of compensatory mobility of the individual in income distribution.

SLIDE 3

3

Hirschman (1973), Ravalion e Lokshin (1999) e Jarvis e Jenkins (1998) suggest that high levels of inequality can or shall be tolerated in societies where the perception of mobility is desirable and possible. However, cases in which high levels of inequality interact with low levels of mobility tend to cause losses of population wellbeing. Gacitua-Marió e Woolcock (2005) highlights that, even in countries that are relatively similar by economic standards, severe barriers to mobility, either social or of other kinds, might concur to exclude certain groups from opportunities access in the job market. Therefore, it becomes possible to measure the length in which long term income, perceived as the mean of the period in consideration, is more or less equally distributed than the income in a given point in this same period. Therefore it better reflects the measurement of population wellbeing. Shorrocks (1978) enhances that mobility is related to levels under which equalization occurs when a period of time is extended. Accordingly, it might be seen as an important society feature regarding the expansion of

• pportunities in the labor market, in which a higher income mobility might result its

bigger convergence, and thus in an improvement of the society income distribution throughout time. Furthermore, the income distribution might present a misleading portrait upon long term inequality, since individual qualification may alter their position in long term income

• distribution. In the case of young individuals, for example, who tend to be located at the

base of the distribution, as a result of low accumulation of human capital, future gains

• f experience throughout the years might positively impact their professional career

ascension, which alters the long term disposition. In case mobility affects initially solely specific groups deprived from a minimum standard of living, albeit incorporating qualification gains or compensatory policies, this may result smaller long-term inequality, and, for a given income inequality, this might express a compensatory effect as a result of ascending mobility for smaller incomes. Therefore, an analysis of the dynamics of income mobility becomes relevant, as it enables measurement levels in which mobility tend to equalize or not the long term income distribution. To measure this behavior, a long wide temporal horizon is necessary. Hence, any

• bserved change in the income distribution might be the outcome of a short period

event that might not consolidate in the long term. According to Solon (2002), an increase in inequality overtime might derive from an increasing differential in the income of the poorer and richer individuals. This sustains a long term increasing

• inequality. Amongst the factors that explain the cross section inequality, the author

shows that any difference between individual features of the same cohort might lead to a permanent variation of income, thus expanding inequality in the long run. A higher educational level, for example, might lead a specific cohort to obtain higher incomes. In addition, individuals may endure an increase in their income volatility, albeit not sustained overtime. In that case, long term inequality might be little affected.

SLIDE 4

4

If on the one hand, mobility can be seen as an indicator of the level of equality of

• pportunities in the labor market as it refers to workers relative income changes
• vertime, on the other hand, it can also be seen as a synonym for income fluctuations

and, thereby, treated as insecure economic factor. Either way, income mobility completes the analysis of income inequality, once mobility may alter, considerably, the long term distribution. Mobility may take place within different aspects. It might be considered in the scope of the individual, known as intragenerational, in which the same individual is observed in two points of time, and, in an intergenerational context, whose unity is the family. The relation of intergenerational income transmission is introduced by the Becker e Tomes (1979) approach. The authors distinguish two courses of action: investment in children human capital derived from parent’s rational decisions and skills correlations, in which genetic and social factors such as IQ transmission, social network and preferences are

• considered. In this line of analysis, the intergenerational transmission of economic

status is identified as an indicator of opportunities of the labor market. When considering the role of education, an opportunity access indicator is generated, given individuals with higher educational levels tend to have greater capacity in finding a workplace, obtain higher wages within a certain occupation and, still, obtain higher chances of career ascension within their occupational status. Bherman et al (2001) highlight that parents education and family economic status are the more commonly used indicators of opportunity access to the labor market. In this context, besides that intergenerational transmission of economic status is identified as an indicator of labor market opportunities, it also tends to be directly related to income inequality overtime. Hence, such as sociological literature emphasizing the social hierarchy, works concerning mobility are based in intergenerational income correlation and educational standards between parents and sons, and are motivated by the theoretical approach of Becker e Tomes (1979, 1986). This matter started to develop within economic literature in the 1990s, stimulated by availability of data in international panels. Björklund e Jäntti (2000), Corak (2004), Aydemir et al (2005) e Erikson e Goldthorpe (2002) are some of the authors that approach this issue. Authors such as Fields and Ok (1996), Solon (1999), Behrman et al (2001) e Ermish e Nicoletti (2005) disclosed the role of intergenerational income elasticity between generations. Due to the limitation of panel

• rganized data, some authors like Ermisch e Francesconi (2004) estimate the

intergenerational elasticity by means of an occupational prestige score - Hope- Goldthorpe score – for parents and sons, ascertained that those are strongly related to the individual income. Within intragenerational scope, variable macroeconomic conditions reflect directly on individual income. Economic growth, for example, may lead to an ascending mobility, whereas stagnation periods generates an opposite behavior. Furthermore, income transfer policies, whenever focused on the poorer population, might impact mobility.

SLIDE 5

5

Jobs being the major source of the population income, the way that workers are hired should have an important impact upon income behavior overtime and, as a consequence, upon the evolution of income distribution. A strong correlation between income mobility and occupational mobility is perceived, since higher occupational status tend to be accompanied by higher incomes than occupational positions of individuals placed in the basis of the social pyramid. In addition, it’s important to highlight the existence of wage differentials, imparting directly on behavior income overtime. The wage difference might exist as a form of compensation of non-pecuniary features amongst work positions occupied by equally potential individuals and by individuals’ diversity, reflected on their productive skills. Further, different earnings may take place for workers that, initially, are equally productive, without any explicit criteria or based

• n non-productive skills. Thereby, these variables might also reflect in bigger or smaller

levels of individuals mobility income. Hence the aim to consider the relation between mobility and inequality regarding the ascending income movement that affects a given segment of the population. For income distribution to be less deconcentrated, gains for those in the basis of the pyramid are required and/or a declining movement for higher wage levels. Material and Methods Empirically, the data from the Monthly Employment Survey (PME) from the Brazilian Institute of Geography and Statistics (IBGE) between 2002 and 2016 is considered. The PME is a research organized longitudinally, albeit undertaking as a rotating panel

• survey. The individual’s information reported on the PME allows measuring mobility in

an intragenerational context in six metropolitan areas- Recife, Salvador, Belo Horizonte, São Paulo, Rio de Janeiro e Porto Alegre – composing the research. The rotating panel allows accompanying households for four straight months and, after an eight months interval, research resumes for another four month period, when they are definitely excluded from the sample. This rotational scheme ensures the longitudinal trait of the research. The sample is constituted by workers, aged between 25 and 65, that presented positive hourly earnings of the main job in the first and fifth interview. From this sample, traditional mobility parameters are calculated, demonstrating immobility, up and descending mobility, and furthermore quantile mobility matrix, displaying motion throughout the quantile of revenue distribution between two points in time. Parameters based in Fields and Ok (1996) are also considered, namely, directional and non-directional mobility income, the movement participation and the mobility as an equalizer instrument of long term incomes. The directional income movement (mdr) is given by





 

n i i i

y x n y x mdr

1

) log (log 1 ) , ( , in which xi and yi represent, respectively, the individual income i on the starting and final

• period. This measure combines income gains and losses, while the effect on trades

SLIDE 6

6

between individuals can be considered by the non-directional income movement (mndr), defined as





 

n i i i

y x n y x mndr

1

| log log | 1 ) , ( . This indicator is decomposed in two effects: distributive, in which mobility may result from a possible alteration in the individuals position inside the income distribution and, economic performance, that reflects the mobility relation with growing periods or economic stagnation. It can be represented by

 

) ) , ( ( ) , ( ) , ( y x mdr mndr y x mdr y x mndr    , in which the first portion is explained by growth, and the second for possible changes in the distribution position. The share movement (mpr) is also estimated and allows measuring the participation of everyone in relation to the total income mean, defined as

n y y x x y x mpr

n i i i





 

1

) , ( , in which x and y represent, respectively, the total income mean on the starting and final period and n, the total number of individuals in the sample. Finally, an emphasis towards mobility as an equalizer of longer-term income is

• considered. The Fields (2005) index is used, once it compares current income

distribution with distribution of the base year, instead of a relative hypothetical path, as it is used in Chakravarty, Dutta e Weimark (1985) and Shorrock´s (1978). If average income is distributed more (or less) equally than the initial income, mobility tends to equalize (or not) the long term income respective to income reference. The index, expressed as )) ( / ) ( ( 1

1 

 

t t

y I y I E , in which yt is the average income in actual period, yt-1 is the base year income vector and I(.) is the measurement of inequality, namely the Gini index. On the one hand, E<0 is used to represent when incomes are unevenly distributed, regarding the base period, throughout time. On the other hand, E>0 is used when it is more evenly distributed. In addition to the indicators, regressions of Least Square Method were considered to measure the possible contribution of different variables for the present income level, and the quantiles regressions for 10o, 25o, 50o, 70o and 90o from the distribution of income based in Koenker (2005) and Koenker and Hallock (2001). Given that a quantile regression allows an analysis of quartiles of income distribution, it enables the observation of which variables have affected income the most, for each of the five income stratus considered. However, the models obtained through the Least Square Method enable an analysis of the average floating income evolution, constituting a comparison parameter with the results obtained for the quantile income.

SLIDE 7

7

Koenker (2005) highlights that linear autoregressive models express the conditional quantile function of the dependent variable as a linear function of the lags of this variable. In this case, both unconditional and conditional mobility can be measured and the model can be expressed as it follows: ) ( ) log( ) ( ) ( ) / (

, 2 1 , 1 1 ) log(

,

i t i t i t Y

v X Y F Q

t i

   

 

       in which ) log(

,t i

Y , is the log income in t, ) log(

1 ,  t i

Y , the lag period,  , the quantile to be estimated,

t i

X , , explanatory variables and, vi, the idiosyncratic error. The independent variables considered are: income in the previous period, age range, educational groups, household conditions, metropolitan region (São Paulo, Rio de Janeiro, Belo Horizonte, Salvador, Porto Alegre and Recife), annual dummies (2003- 2015), and, further, socio-occupational status groups, based on the rating proposed by Monsueto (2016), that considers aspects of productivity and quality of work posts. Classes are divided into: high, high-medium, medium, medium-low and low. Results Distinct trends of employment income evolution are observed with respect to the groups analyzed (Figures 1 and 2). Greater wage dispersion can be seen for whites, in both periods studied, with reduction in the recent period. The non-white presents smaller wage dispersion, accentuated in the recent period, for smaller wage levels. Figure 1- Log of hourly earnings distribution of work by color, women for metropolitan regions in Brazil between 2005 and 2015

Source: PME, 2002-2015

,2 ,4 ,6 ,8 1

• 2

2 4 6

log of hourly earnings

Non-white women - 2005 White women - 2005 Non-white women - 2015 White women - 2015

SLIDE 8

8

Figure 2 - Log of hourly earnings distribution of work by color, men for metropolitan regions in Brazil between 2005 and 2015

Source: PME, 2002-2015

According to mobility indicators, it is noticeable if gains in the period reflect solely conjuncture changes and/or are accompanied by alterations of the individuals position in the income distribution (Table 1). Through directional income movement, it is perceivable that a positive variation in the individual average gains are more significant for women and non-white. From 2012 onwards, the indicator presents a declining tendency and the difference between groups is reduced. This indicator remains in smaller levels than non-directional movement, attaining little change throughout the

• period. By decomposing this indicator, a bigger distribution effect is evidenced, mainly

in the recent period when the economic effect reaches levels close to zero. This behavior is confirmed by the share movement, which reaches a prominent position due to its bigger magnitude between the considered indicators. Albeit with decreasing tendency for all groups, this indicator stands in a higher level than the USA and France, as reported by Fields et al (2000) e Buchinsky et al (2003), for the USA (between 1970 and 1995) and for France (from 1978), respectively. Through the equalization of longer- term incomes, close values to the French case and below those American numbers are

• bserved. Fields (2005) points values between 0.008 and 0.004 for the United States,

between 1970 and 1995. Buchinsky et al (2003) show values between 0.04 and 0.021 for France. Amongst groups, non-whites retain the higher values for the whole period. Mobility indicates average gains for the groups in that period. However, it does not allow an analysis about individuals that benefited with these movements throughout

• distribution. The mobility direction, ascending mobility, immobility, descending

mobility and transition matrix assist to further these movements.

,2 ,4 ,6 ,8

• 2

2 4 6

log of hourly earnings Non-white men - 2005 White men - 2005 Non-white men - 2015 White men - 2015

SLIDE 9

9

Table 1- Mobility indicators, 2003-2015

Source: PME, 2002-2015

Figure 3 - Descending and ascending mobility and immobility, Metropolitan Areas in Brazil, 2003-2015

Source: PME, 2002-2015

2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 Directional movement Non-white women 0,11 0,09 0,12 0,10 0,10 0,12 0,09 0,10 0,10 0,12 0,11 0,08 0,06 White women 0,08 0,06 0,08 0,09 0,09 0,10 0,08 0,08 0,08 0,09 0,07 0,08 0,05 Non-white men 0,10 0,07 0,10 0,10 0,09 0,10 0,08 0,11 0,09 0,11 0,08 0,08 0,05 White men 0,07 0,09 0,09 0,10 0,09 0,10 0,07 0,10 0,07 0,09 0,07 0,07 0,04 Non-directional movement Non-white women 0,48 0,36 0,36 0,33 0,31 0,31 0,29 0,31 0,31 0,31 0,31 0,32 0,32 White women 0,44 0,39 0,34 0,34 0,31 0,31 0,31 0,30 0,33 0,32 0,30 0,32 0,32 Non-white men 0,45 0,36 0,34 0,30 0,29 0,29 0,28 0,29 0,29 0,30 0,31 0,32 0,32 White men 0,44 0,37 0,34 0,33 0,30 0,31 0,31 0,31 0,33 0,33 0,33 0,32 0,33 total 0,45 0,37 0,34 0,33 0,30 0,30 0,30 0,30 0,32 0,31 0,31 0,32 0,32 Economic effect Non-white women 0,11 0,09 0,12 0,10 0,10 0,12 0,09 0,10 0,10 0,12 0,11 0,08 0,06 White women 0,08 0,06 0,08 0,09 0,09 0,10 0,08 0,08 0,08 0,09 0,07 0,08 0,05 Non-white men 0,10 0,07 0,10 0,10 0,09 0,10 0,08 0,11 0,09 0,11 0,08 0,08 0,05 White men 0,07 0,09 0,09 0,10 0,09 0,10 0,07 0,10 0,07 0,09 0,07 0,07 0,04 Distributive effect Non-white women 0,37 0,27 0,24 0,23 0,21 0,20 0,20 0,21 0,21 0,19 0,20 0,23 0,26 White women 0,36 0,33 0,26 0,26 0,22 0,21 0,22 0,22 0,25 0,23 0,23 0,23 0,27 Non-white men 0,36 0,29 0,23 0,21 0,20 0,19 0,20 0,19 0,21 0,19 0,23 0,24 0,28 White men 0,37 0,28 0,25 0,23 0,21 0,21 0,23 0,22 0,25 0,24 0,25 0,25 0,29 Share movement Non-white women 0,50 0,36 0,36 0,35 0,31 0,32 0,31 0,34 0,33 0,32 0,32 0,34 0,35 White women 0,43 0,41 0,36 0,33 0,30 0,30 0,32 0,30 0,36 0,33 0,33 0,33 0,35 Non-white men 0,45 0,35 0,34 0,31 0,30 0,31 0,29 0,31 0,32 0,32 0,34 0,36 0,36 White men 0,44 0,39 0,36 0,34 0,30 0,31 0,32 0,34 0,37 0,36 0,40 0,35 0,36 Equalizer of longer-term income (Gini Index) Non-white women 0,05 0,05 0,05 0,03 0,02 0,02 0,04 0,03 0,03 0,00 0,02 0,06 0,02 White women 0,01 0,04 0,02 0,03 0,02 0,02 0,02 0,03 0,01 0,02 0,02 0,01 0,03 Non-white men

• 0,01

0,02 0,03 0,01 0,05 0,03 0,04 0,01 0,04 0,04 0,04 0,04 0,05 White men 0,00 0,03 0,01 0,02 0,02 0,01 0,03 0,02 0,02 0,00

• 0,01

0,02 0,03

SLIDE 10

10

Given the figure 3, it is evident that immobility is higher for white men, followed by white women. The non-whites register higher levels of mobility, ascending and descending. The Gini index illustrates inequality between comparison groups. The downward movement shows average gains for all groups, albeit with higher inequality levels for white men, followed by white women. Non-white men and women have lower values. Less inequality along with transition matrix analysis allows a better portrait of inequality distribution. Non-white men and women attain higher proportion amongst smaller levels of income, reflected in smaller inequality indexes. On the other hand, both white men and women are concentrated in the higher levels of income distribution. Figure 4 – Gini Index for groups of comparison, Metropolitan Region in Brazil, 2003- 2015

Source: PME, 2003-2015

The frequency distribution of transition matrix for the years of 2005 and 2015 displays that white men and women concentrate higher immobility in higher deciles of wage

• distribution. The 20% richer concentrate around 21% of white women total and 25% of

white men, whereas non-white women register 6% and non-white men 10% in recent

• period. The analysis of the 40% in the top of the distribution sustain those differences.

An opposite composition is seen when immobility of the 20% poorer in wage terms and 40% poorer is observed. In this context, non-white women register the higher percentages, followed by non-white men. Ascending mobility at the base and descending at the top may lead to higher income equalization overtime, albeit in level inequality remains high. Even though intragenerational mobility has been low in the period, if sustained for prolonged periods, it might lead to lower long term inequality.

SLIDE 11

11

Table 2 - Quantile mobility matrix, Metropolitan Areas in Brazil, 2004/2005 and 2014/2015

Non-white women – 2005 Non-white women – 2015 5ª 5ª 1 2 3 4 5 1 2 3 4 5 1 25,30 6,05 2,92 0,73 0,12 1 21,32 6,94 2,80 1,04 0,22 2 7,14 11,94 4,19 1,00 0,18 2 6,37 13,09 4,30 1,40 0,35 1ª 3 1,49 4,59 10,58 2,12 0,39 1ª 3 2,16 4,37 8,20 3,00 0,81 4 0,54 1,00 2,64 7,94 1,47 4 1,34 1,54 2,07 6,83 2,10 5 0,08 0,24 0,55 1,76 5,04 5 0,21 0,37 0,59 1,95 6,64 White women – 2005 White women – 2015 5ª 5ª 1 2 3 4 5 1 2 3 4 5 1 10,09 3,75 1,16 0,48 0,13 1 9,28 3,99 1,61 0,68 0,31 2 4,20 8,85 3,86 1,17 0,19 2 3,82 9,81 3,64 1,31 0,40 1ª 3 1,37 3,96 7,83 2,89 0,86 1ª 3 1,38 3,36 8,49 3,32 1,03 4 0,46 1,53 3,03 11,89 3,92 4 0,62 1,51 3,14 11,63 3,68 5 0,14 0,26 0,99 4,33 22,65 5 0,29 0,39 1,13 3,72 21,47 Non-white men – 2005 Non-white men – 2015 5ª 5ª 1 2 3 4 5 1 2 3 4 5 1 15,08 4,87 1,49 0,79 0,13 1 10,22 4,07 2,13 0,86 0,20 2 5,17 12,14 4,24 1,38 0,25 2 4,23 10,43 5,04 1,91 0,48 1ª 3 1,77 5,18 12,75 3,57 0,63 1ª 3 2,17 4,92 12,43 4,41 0,95 4 0,51 1,44 4,34 11,49 2,24 4 1,01 2,00 4,40 11,66 2,34 5 0,22 0,26 0,51 2,00 7,56 5 0,24 0,40 1,07 2,39 10,05 White men – 2005 White men – 2015 5ª 5ª 1 2 3 4 5 1 2 3 4 5 1 4,81 2,47 1,21 0,46 0,12 1 3,88 1,86 1,54 0,66 0,32 2 2,74 6,84 3,68 1,40 0,28 2 2,02 6,75 3,26 1,55 0,43 1ª 3 0,93 3,44 10,71 4,32 0,95 1ª 3 1,00 3,27 10,14 4,55 1,10 4 0,48 1,15 4,09 14,03 4,32 4 0,63 1,59 4,64 14,43 4,75 5 0,08 0,38 0,77 4,28 26,05 5 0,26 0,37 1,31 4,33 25,36

Source: PME, 2004/2005 e 2014/2015

The results obtained by the estimation of quantile regressions indicate higher mobility for non-whites in Brazil in the beginning of XXI century, although with differential

• persistence2. In line with this result, Osório (2009), for example, shows less social

ascension for black and, according to the author, the persistence of wage income inequality is mostly due to educational disadvantages between racial groups. This is also highlighted by Soares (2000) and Machado, Wajnman and Oliveira (2006).

2 All the estimates are in the Annex.

SLIDE 12

12

The estimation for intragenerational mobility can be seen in the figure below. The closer the coefficients for income variable in t-1 are to one, the higher the temporal

• dependency. Hence, as income in the previous period is a condition to income

distribution, the results close to one indicates a higher level of immobility between

• periods. Estimation differences displays the existence of different mobility patterns for

the four comparison groups. Income mobility decreases until the median, when it starts to increase, with higher expression for the poorer and richer decile of the population. Non-white women register higher mobility, followed by non-white men in all quantiles

• f income distribution, with the exception of the poorer decile, in which non-white men

have more expression. Figure 5 - Intertemporal dependency conditioned by comparison groups, according to quantiles

Source: PME, 2002-2015

The coefficients of education, measured by groups divided in years of study, show non- homogeneity throughout distribution. The educational yield is increasing for all of those that are above the median in the distribution throughout the period. Similar evidences are documented in Maciel et al (2001). When compared with international findings, the role of education has more significance for the Brazilian low-income population, given the smaller intertemporal dependency of income at the base of the distribution. Navarro (2006) points out similar results, although less expressive for the Argentinian case. Buchinsky (2001) shows that, in the case of the U.S, the educational yield is higher for lower income quantiles at the beginning of the period (1968) and higher in the upper quantiles by the end of the period (1990). In this sense, the educational level of the Brazilian population may constitute an important element for income mobility and thus less concentration of the distribution, leading to an income reversion through time in favor of the less fortunate. This income reversion is observed by Figueiredo et al (2007), when considering the role of education on wage return.

SLIDE 13

13

In addition, being in a higher occupational class has a positive impact in the workers income throughout wage distribution, mainly for non-white women. Therefore, it becomes more expressive for higher wage levels. Amongst other individual traits, being the household chief attains higher importance for those located at the inferior tail of distribution, especially, for men. Yet, the age variable has a positive impact on the wage return, mainly for those located at the higher deciles. A more distinctive behavior for men than women, regardless of color. Institutional and global factors might justify an increase in mobility. These factors are measured indirectly by annual dummies. The coefficients obtained through the Least Square Method have shown a tendency of growth in average work income in the whole country throughout the first decade of the XXI century. It is noticeable a positive effect for all quantiles, being more expressive for men. In conclusion, the impact metropolitan living areas on income was small in comparison to other variables. Final Considerations A mobility analysis in a intragenerational context was possible due to individual data

• ffered by the Monthly Employment Survey between March 2002 and February 2016,

which was the last month available. Mobility indicators and quantile regressions results have shown that mobility contributes for a more deconcentrated distribution throughout time, in different levels for white and non-white. The non-white has registered more mobility, mostly for women, along with lower levels of inequality and higher concentration at the basis of the wage pyramid. Further, higher levels of ascending mobility for non-white at the basis of distribution might contribute for lower inequality between groups in the long

• term. Individual factors, mainly schooling, has great importance for ascending mobility

at the basis of the pyramid, albeit tends to contribute for the maintenance of high inequality levels between those with higher incomes. Income distribution in Brazil still portraits one of the highest levels of iniquity in the

• world. Nevertheless, some progress in this scenario was evidenced in this study. The

improvement of public policies aimed at social inclusion may assure the persistence of income mobility at the basis of distribution, contributing, in the near future, for wellbeing gains for the population, and especially its poorest part. References AYDEMIR, Abdurrahman; CHEN, Wen-Hao; CORAK, Miles. Intergenerational earnings mobility among the children of Canadian immigrants. Analytical Studies Branch Research Paper Series 2005267, Statistics Canada, Analytical Studies Branch, 2005.

SLIDE 14

14

BECKER, G. S.; TOMES, N. An Equilibrium Theory of the Distribution of Income and Intergenerational Mobility, Journal of Political Economy, 87, p. 1153–1189, 1979. BECKER, G. S.; TOMES, N. Human Capital and the Rise and Fall of Families, Journal of Labor Economics, 4, p. 1–39, 1986. BEHRMAN, J. R.; GAVIERIA, A. U.; SZEKELY, M. S. Intergenerational Mobility in Latin America. Inter-American Development Bank, Research Department Working Papers Series, n. 452, 2001. BJÖRKLUND, A.; JÄNTTI, M. Intergenerational Mobility of Socio-economic Status in Comparative Perspective, Nordic Journal of Political Economy, 26, p. 3–32, 2000. BUCHINSKY, M. Quantile regression with sample selection: Estimating women’s return to education in the U.S., Empirical Economics, 26: 87-113, 2001. BUCHINSKY, Moshe; FIELDS, Gary; FOUGÈRE, Denis; KRAMARZ, Francis. Francs or Ranks? Earnings Mobility in France, 1967-1999. London: Centre for Economic Policy Research, C.E.P.R. Discussion Papers, n. 3937, June 2003. BUCHINSKY, Moshe; HUNT, Jennifer. Wage mobility in the United States. The Review of Economics and Statistics, vol.LXXXI, n.3, p.351-368, August 1999. CHAKRAVARTY, S.J.; DUTTA, B.; WEYMARK, J.A. Ethical indices of income

• mobility. Social Choice and Welfare, Berlin, v. 2, n. 1, p. 1-21, May. 1985.

CORAK, M. Do poor children become poor adults? Lessons for public policy from a cross country comparison of generational earnings mobility. Paper presented at Colloque surle devenir des enfants de familles défavorisées en France, April, 2004. ERMISCH J.; NICOLETTI, C. Intergenerational earnings mobility: Changes across cohorts in Britain, ISER Working Paper 2005-19. Colchester: University of Essex, 2005. ERIKSON, R.; GOLDTHORPE, J.H. Intergenerational inequality: a sociological

• perspective. Journal of Economic Perspective, 16, p. 31-44, 2002.

ERMISCH, J.; FRANCESCONI, J. Intergenerational Mobility in Britain: New Evidence from the British Household Panel Study. In: CORAK, M. (editor) Generational Income Mobility in North America and Europe. New York: Cambridge University Press, 2004. FIELDS, Gary. Does income mobility equalize longer-term incomes? New measures

• f an old concept. New York: Cornell University, 2005. Paper presented on Frontiers
• n Social and Economic Mobility Conference.

Fields, Gary S., and Efe A. Ok. "The meaning and measurement of income mobility." Journal of Economic Theory 71.2 (1996): 349-377. FIELDS, Gary; LEARY, Jesse; OK, Efe. Dollars and Deciles: Changing Earnings Mobility in the United States, 1970-1995. Cornell University working paper, 2000. FIGUEIREDO, E.; NETTO JUNIOR, J.; PORTO JUNIOR, S. Distribuição, mobilidade e polarização de renda no Brasil: 1987 a 2003. Revista Brasileira de Economia, n.61, pp.1-27, 2007 GACITUA-MARIÓ, Estanislao; WOOLCOCK, Michael. Uma avaliação da exclusão social e da mobilidade no Brasil. In: GACITUA-MARIÓ, Estanislao; WOOLCOCK, Michael (Org.) Exclusão social e mobilidade no Brasil. Brasília: IPEA, 2005

SLIDE 15

15

GITTLEMAN, M.; JOYCE, M. Earnings Mobility and Long-Run Inequality: An Analysis Using Matched CPS Data. Industrial Relations, 35, p.180-196, 1996. GOTTSCHALK, P.; MOFFIT, R. The Growth of Earnings Instability in the U.S. Labor

• Market. Brookings Papers on Economic Activity, 2, p.217-272, 1994.

HIRSCHMAN, A. The changing tolerance for income inequality in the course of economic development, with a mathematical appendix by Michael Rothschild. Quartely journal of economics, vol.87, p. 544-566, 1973

• IBGE. Série histórica da pesquisa mensal de emprego. Disponível em

<www.ibge.gov.br> JARVIS, S.; JENKINS, S.P. How much income mobility is there in Britain? The Economic Journal, 108, p.428-443, 1998. Koenker, Roger. Quantile regression. No. 38. Cambridge university press, 2005. Koenker, Roger, and Kevin Hallock. "Quantile regression: An introduction."Journal of Economic Perspectives 15.4 (2001): 43-56. MACHADO, A. F.; OLIVEIRA, A. M. H. C.; WAJNMAN, S. Sexo Frágil? Evidências sobre a inserção da mulher no mercado de trabalho brasileiro. São Paulo: Organização Gelre, 2005. v. 1. 68 p. MACIEL, M.C.; CAMPÊLO, A.K.; RAPOSO, M.C.F. A Dinâmica das Mudanças na Distribuição Salarial e no Retorno em Educação para Mulheres: uma aplicação de regressão quantílica. In: Anais do Encontro ANPEC, 2001. MONSUETO, Sandro Eduardo. NOTA TÉCNICA EM ECONOMIA n. 07 NAVARRO, A.I. Estimating income mobility in Argentina with pseudo-panel data. Paper presented at LACEA, 2006. OSORIO, R. G. Mobilidade social sob a perspectiva da distribuição de renda. 2003. Dissertação (Mestrado) – Instituto de Ciências Sociais, Departamento de Sociologia, Universidade de Brasília, Brasília. RAVALION, M.; LOKSHIN, M. Who wants to redistribute? Russia`s tunnel effect in

• 1990s. Policy research working paper series n.2150. Washington, DC: The World

Bank, 1999. SHORROCKS, A.F. The measurement of mobility. Econometrica, v.46, n. 5, p. 1013- 24, 1978. SOARES, S. S. D. O perfil da discriminação no mercado de trabalho: homens negros, mulheres brancas e mulheres negras. Brasília: Ipea, Texto para Discussão, n. 769, 2000. SOLON, G. Intergenerational Mobility in the Labor Market. In: Orley C. Ashenfelter and David Card (editors). Handbook of Labor Economics, Volume 3A, Amsterdam: Elsevier Science, 1999. SOLON, G. Cross-country Differences in Intergenerational Earnings Mobility, Journal

• f Economic Perspectives, 16, p. 59–66, 2002.

SLIDE 16

16 Annex Table A1 – Conditioned Mobility, Metropolitan Areas in Brazil, 2003-2016

Source: PME, 2002-2015

MQO 0,10 0,25 0,50 0,75 0,90 MQO 0,10 0,25 0,50 0,75 0,90 Log of hourly earnings in t-1 0,616 0,568 0,691 0,768 0,700 0,606 0,660 0,581 0,737 0,820 0,751 0,643 4 to 7 years of schooling 0,042 0,098 0,041 0,012 0,024 0,031 0,048 0,083 0,031 0,017 0,026 0,043 8 to 10 years of schooling 0,087 0,130 0,066 0,032 0,053 0,068 0,097 0,147 0,068 0,039 0,070 0,092 11 years of schooling 0,175 0,232 0,113 0,077 0,129 0,188 0,194 0,234 0,116 0,084 0,151 0,232 12 or more years of schooling 0,408 0,379 0,262 0,215 0,350 0,474 0,449 0,462 0,280 0,218 0,356 0,541 Spouse

• 0,004

0,004

• 0,003
• 0,009
• 0,005
• 0,016
• 0,007
• 0,005
• 0,007
• 0,012
• 0,008
• 0,001

Son

• 0,042
• 0,016
• 0,022
• 0,025
• 0,033
• 0,066
• 0,069
• 0,080
• 0,050
• 0,038
• 0,044
• 0,056

Other parent

• 0,041

0,022 0,000

• 0,010
• 0,041
• 0,068
• 0,041
• 0,032
• 0,020
• 0,023
• 0,035
• 0,062

Age - 35 to 44 years 0,029 0,008 0,008 0,011 0,026 0,048 0,027 0,006 0,016 0,013 0,024 0,045 Age - 45 to 54 years 0,047

• 0,006

0,010 0,020 0,049 0,099 0,049

• 0,006

0,010 0,022 0,056 0,103 Age - 55 to 65 years 0,063

• 0,047

0,005 0,029 0,077 0,150 0,058

• 0,038

0,006 0,026 0,075 0,157 Recife - metropolitan area

• 0,103
• 0,259
• 0,133
• 0,046
• 0,022

0,045

• 0,091
• 0,279
• 0,144
• 0,018

0,024 0,054 Salvador - metropolitan area

• 0,069
• 0,187
• 0,084
• 0,032
• 0,010

0,035

• 0,014
• 0,094
• 0,052
• 0,005

0,047 0,076 Belo Horizonte - metropolitan area 0,003

• 0,063
• 0,044

0,005 0,043 0,073

• 0,006
• 0,088
• 0,056

0,008 0,059 0,087 Rio de Janeiro - metropolitan area

• 0,028
• 0,010
• 0,008
• 0,020
• 0,053
• 0,053
• 0,043
• 0,036
• 0,014
• 0,026
• 0,056
• 0,065

Porto Alegre - metropolitan area 0,005

• 0,065
• 0,038

0,012 0,048 0,087

• 0,018
• 0,060
• 0,041

0,003 0,022 0,023 Occupational class - high 0,121 0,048 0,057 0,083 0,135 0,204 0,138 0,106 0,094 0,083 0,126 0,185 Occupational class - medium/high 0,043 0,042 0,030 0,028 0,047 0,051 0,051 0,052 0,037 0,026 0,042 0,062 Occupational class - medium/high

• 0,006

0,048 0,013

• 0,006
• 0,022
• 0,048

0,005 0,028 0,012 0,003

• 0,007
• 0,020

Occupational class -medium/low

• 0,011
• 0,162
• 0,077
• 0,002

0,068 0,135 0,020

• 0,111
• 0,035

0,011 0,079 0,154 2004 0,003 0,149 0,073

• 0,001
• 0,039
• 0,072

0,024 0,052 0,046 0,021 0,019 0,032 2005 0,039 0,212 0,104 0,021

• 0,016
• 0,077

0,054 0,148 0,101 0,038 0,000

• 0,024

2006 0,058 0,276 0,140 0,041

• 0,011
• 0,063

0,087 0,212 0,136 0,051 0,020 0,000 2007 0,094 0,331 0,178 0,051

• 0,004
• 0,063

0,108 0,279 0,163 0,059 0,016

• 0,007

2008 0,140 0,383 0,207 0,075 0,040

• 0,015

0,137 0,309 0,187 0,080 0,039

• 0,002

2009 0,141 0,400 0,216 0,084 0,031

• 0,008

0,146 0,328 0,189 0,089 0,059 0,019 2010 0,174 0,423 0,254 0,104 0,052 0,025 0,176 0,364 0,218 0,098 0,061 0,047 2011 0,225 0,508 0,288 0,124 0,082 0,068 0,188 0,359 0,216 0,102 0,079 0,114 2012 0,269 0,529 0,320 0,151 0,132 0,144 0,231 0,448 0,264 0,129 0,107 0,109 2013 0,295 0,566 0,332 0,162 0,153 0,161 0,255 0,477 0,281 0,131 0,117 0,139 2014 0,320 0,605 0,356 0,184 0,171 0,185 0,274 0,511 0,295 0,137 0,125 0,162 2015 0,333 0,625 0,373 0,190 0,171 0,188 0,272 0,501 0,296 0,138 0,129 0,161 2016 0,352 0,655 0,393 0,193 0,177 0,239 0,285 0,476 0,322 0,153 0,133 0,141 Constant 0,282

• 0,273

0,026 0,226 0,471 0,799 0,255

• 0,192

0,015 0,177 0,412 0,740 Non-white women White women MQO 0,10 0,25 0,50 0,75 0,90 MQO 0,10 0,25 0,50 0,75 0,90 Log of hourly earnings in t-1 0,628 0,538 0,705 0,789 0,727 0,630 0,661 0,570 0,751 0,828 0,745 0,634 4 to 7 years of schooling 0,058 0,083 0,039 0,033 0,039 0,055 0,075 0,088 0,027 0,038 0,064 0,103 8 to 10 years of schooling 0,111 0,144 0,074 0,058 0,084 0,121 0,123 0,129 0,048 0,059 0,105 0,167 11 years of schooling 0,183 0,196 0,112 0,098 0,144 0,219 0,220 0,199 0,111 0,107 0,195 0,309 12 or more years of schooling 0,389 0,360 0,244 0,214 0,318 0,486 0,449 0,418 0,255 0,226 0,380 0,599 Spouse

• 0,015

0,002

• 0,010
• 0,005
• 0,016
• 0,026
• 0,019
• 0,028
• 0,018
• 0,011
• 0,008
• 0,005

Son

• 0,070
• 0,090
• 0,048
• 0,033
• 0,034
• 0,045
• 0,101
• 0,135
• 0,080
• 0,050
• 0,058
• 0,077

Other parent

• 0,070
• 0,046
• 0,040
• 0,039
• 0,056
• 0,104
• 0,079
• 0,096
• 0,064
• 0,039
• 0,055
• 0,070

Age - 35 to 44 years 0,024 0,001 0,016 0,018 0,029 0,052 0,041 0,008 0,016 0,014 0,043 0,090 Age - 45 to 54 years 0,051 0,007 0,025 0,031 0,059 0,111 0,064 0,002 0,014 0,025 0,077 0,147 Age - 55 to 65 years 0,049

• 0,030

0,012 0,028 0,072 0,152 0,075

• 0,035

0,001 0,031 0,103 0,200 Recife - metropolitan area

• 0,098
• 0,230
• 0,137
• 0,046

0,001 0,011

• 0,086
• 0,235
• 0,141
• 0,033

0,002 0,049 Salvador - metropolitan area

• 0,064
• 0,155
• 0,084
• 0,041
• 0,014

0,019

• 0,016
• 0,103
• 0,041
• 0,005

0,024 0,065 Belo Horizonte - metropolitan area 0,013

• 0,060
• 0,037

0,016 0,061 0,077 0,014

• 0,055
• 0,039

0,026 0,069 0,068 Rio de Janeiro - metropolitan area

• 0,044
• 0,009
• 0,007
• 0,032
• 0,075
• 0,112
• 0,043
• 0,020
• 0,006
• 0,029
• 0,063
• 0,095

Porto Alegre - metropolitan area 0,001

• 0,060
• 0,041

0,017 0,041 0,057

• 0,016
• 0,060
• 0,040

0,003 0,026 0,012 Occupational class - high 0,134 0,045 0,051 0,080 0,136 0,199 0,169 0,144 0,115 0,088 0,149 0,216 Occupational class - medium/high 0,056 0,071 0,040 0,026 0,034 0,052 0,058 0,074 0,048 0,029 0,039 0,058 Occupational class - medium/high 0,012 0,051 0,020 0,000

• 0,014
• 0,029

0,027 0,064 0,030 0,013 0,008

• 0,009

Occupational class -medium/low

• 0,010
• 0,112
• 0,061
• 0,006

0,033 0,109 0,049

• 0,020
• 0,008

0,023 0,082 0,158 2004 0,010 0,114 0,056 0,006

• 0,043
• 0,125

0,022 0,091 0,044 0,015

• 0,011
• 0,051

2005 0,077 0,232 0,132 0,046

• 0,007
• 0,102

0,051 0,146 0,075 0,024

• 0,007
• 0,043

2006 0,097 0,277 0,154 0,052

• 0,012
• 0,116

0,078 0,235 0,112 0,036 0,004

• 0,064

2007 0,115 0,335 0,179 0,058

• 0,010
• 0,103

0,103 0,286 0,135 0,046 0,007

• 0,042

2008 0,159 0,400 0,217 0,086 0,033

• 0,061

0,118 0,295 0,145 0,059 0,023

• 0,032

2009 0,172 0,417 0,233 0,095 0,037

• 0,060

0,129 0,329 0,161 0,064 0,031

• 0,021

2010 0,226 0,481 0,262 0,116 0,071 0,005 0,170 0,376 0,188 0,086 0,059 0,015 2011 0,243 0,503 0,282 0,129 0,080 0,028 0,184 0,396 0,191 0,084 0,074 0,064 2012 0,296 0,562 0,307 0,152 0,121 0,094 0,211 0,439 0,215 0,109 0,099 0,096 2013 0,309 0,572 0,322 0,160 0,136 0,101 0,233 0,446 0,229 0,115 0,115 0,136 2014 0,344 0,625 0,349 0,180 0,155 0,127 0,247 0,500 0,250 0,117 0,102 0,109 2015 0,346 0,631 0,361 0,180 0,148 0,135 0,245 0,508 0,244 0,110 0,104 0,104 2016 0,360 0,658 0,353 0,187 0,189 0,149 0,223 0,526 0,234 0,130 0,093 0,033 Constant 0,319

• 0,162

0,043 0,220 0,508 0,889 0,322

• 0,095

0,078 0,200 0,471 0,833 Non-white men White men