OCCUPATIONAL TRANSITIONS IN THE METROPOLITAN LABOR MARKET - AN - - PDF document

SLIDE 1

1

OCCUPATIONAL TRANSITIONS IN THE METROPOLITAN LABOR MARKET - AN ANALYSIS BY COLOR AND GENDER IN THE YEARS 20001 Sandro Eduardo Monsueto – Universidade Federal de Goiás Mariangela Furlan Antigo – Universidade Federal de Minas Gerais Daniela Goes Paraiso Lacerda – Fundação João Pinheiro Jaqueline Moraes Assis Gouveia - Universidade Estatual de Campinas Abstract This article intends to study occupational transitions with different levels of qualification, by gender and color, in the period between 2002 and 2016. From the data of the Monthly Employment Survey applied to six metropolitan regions of the country, the results show that whites individuals stand out for their greater distribution of time in occupations with high qualification. Assigning the probability pattern of occupational transition from whites to nonwhites shows that the fraction of time in which the latter pass in the best

• ccupations increases, while the time spent in the two categories of occupations of lower

socioeconomic quality is reduced. In addition, the multinomial logit models show a significant temporal dependence of the contemporary occupational situation of the workers, evidencing difficulties to carry out transitions between the segments of socioeconomic quality considered. This result is more expressive to non-white workers. Keywords: transition; occupation class; gender; color. Introduction The successful stabilization plan implemented during the 1990s, the valuation of the minimum wage, the creation of income transfer policies, the generation of employment and the greater formalization of the labor force are some examples of policies that have improved the distribution and conditions of the labor market in Brazil in the first decade

• f the 21st century, with a trend of reversion in the current period. Given these results,

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

questions arise regarding the dynamics of the Brazilian labor market, especially in the view of the recent rise in unemployment and informality rates. In this sense, this article aims to study the Brazilian labor market from both a general point of view and considering the role of sex and color. Among the main objectives of the study, we highlight the search for the variables that explain the probabilities of success

• f the individuals in the search for occupation, the identification of personal attributes

that affect these probabilities and the investigation of which groups (white men, non- white men, white women and non-white women) are more benefited from the behavior

• f the labor market in Brazil in the first decade of the twenty-first century and in the most

recent period. It is also considered the quality of the work place occupied by each for each demographic group analyzed. Theoretical foundation Several studies have highlighted the role of the labor market and the occupational structure both to reflect the social inequalities in the country (Barros and Mendonça, 1995 and Ramos and Vieira, 2001), as well as to generate additional inequalities, given the presence of segmentation and discrimination of gender and race (Oliveira and Ribeiro, 1998; Oliveira, 1998 and 2003; Soares and Oliveira, 2004; Araújo and Ribeiro, 2001; Matos and Machado, 2006). However, the Brazilian labor market can also act as a source

• f exit for situations of very low income or poverty (Barros et al, 1997). That is, the

insertion in the labor market is an instrument by which many workers can reach income levels above the poverty line. Thus, the structure of the market, primarily in terms of

• ccupational segmentation can be an important means of combating or expansion of the

pay gap, as argued in, for example, Amadeo et al (1994). If the occupational structure is important to explain the income gap, so how workers change occupational status (idle, employed or unemployed) or, once employed, how they can go out of certain occupational segment can change the setting of this differential. Examples of this relationship can be obtained by the evidence that workers can use

• ccupational mobility to earn salary increases and as a way out of precarious work

situations, such as informal or high incidence of involuntary unemployment (Contini and Villosio, 2000; Holzer et al, 2003; García Pérez and Sanz, 2005; Davia, 2006).

SLIDE 3

3

On the other hand, segmentation and occupational segregation are phenomena that limit the mobility of workers between status and type of occupation, causing, among other things, a change in occupational status only reproduce the uneven structure of allocation

• f labor throughout time, without signifying improvements in the social condition of the
• individual. Research reveals differences in mobility patterns among population groups,

gender, race, and different levels of income. Overall, it seems that the mobility rate is higher among men than women (Gabriel, 2003; Parrado and Wolf, 1999) and that groups with higher incomes are more stable (Parrado et al , 2007). We also observe some restrictions on occupational mobility towards high yield

• ccupations because this movement would be limited by discrimination, imperfect

information, the disabled employment networks and the preference of agents (Holzer et al., 2003). Paci and Serneels (2007) show that there are significant barriers to

• ccupational mobility between formal and informal segments, determined, among other

factors, by education and access to capital. For the specific case of Brazil, Neri et al (1997) show a clear distinction between mobility behavior for formal and informal workers, which is significantly higher for the latter with less experienced individuals. In addition, the pattern of mobility is differentiated by gender and race, so that women and black Brazilians would be over-represented in low- paying occupations and would present occupational trajectories not always favorable, for example, a descendant excess mobility and a lower salary premium for changing

• ccupation (Pinto and Neri, 2000).

Previous results also show evidence that segmentation of the labor market can generate groups of workers with low experience and specific human capital, with no defined professional future. These workers tend to have greater difficulty in professional relocation and are likely to be more vulnerable to involuntary changes in employment or

• ccupation. The results also seem to indicate the existence of an excess of involuntary

mobility among the poorer workers or between specific demographic groups. Such mobility, according to the view of occupational matches, tends to generate negative impact on pay and on the welfare of individuals (McLaughlin, 1991). Therefore, it is verified that the occupational structure is important to explain the wage

• gap. On the one hand, the way that workers change occupation or how they manage to

leave a occupational segment can change the configuration of the income distribution. On

SLIDE 4

4

the other hand, segmentation and occupational segregation are phenomena that limit the mobility of workers, so that, among other effects, a change of occupation only reproduces the unequal structure of labor allocation over time, without improving the social condition

• f the individual (Fitzenberger and Kuzne, 2005; Maltselva, 2005). That is, wage

inequality can be influenced both by the occupational structure and by the workers' movements that alter this structure. The practical implication of these results is that public policies aimed at wage equalization should not only consider the present occupational situation of workers, but should also consider the level and quality of occupational movements. They must consider not only current wages, but also the possibilities of professional and socio-economic advancements throughout the life cycle. The generation of jobs concentrated in jobs or sectors of low mobility towards occupations of better status can accentuate the existing barriers, far from promoting the homogeneity of opportunities. In this way, it is important to identify which occupations are most used as a means of entry into the labor market or those that are the subject of public employment generation policies, verifying if they are sources of upward or downward mobility. Methodology The database used in this work is the Pesquisa Mensal do Emprego (Monthly Employment Survey - (PME)), conducted by the Instituto Brasileiro de Geografia e Estatística (Brazilian Institute of Geography and Statistics (IBGE)), during the period 2002 to 2016. The sample survey is made for six metropolitan regions of Brazil, namely Recife, Salvador, Belo Horizonte, Rio de Janeiro, São Paulo and Porto Alegre. One of the main advantages of the database generated by this research is the ability to follow up the individual over time, for a period of up to one year. Two methods of analysis are used in this article. The first is a univariate analysis based

• n the work of Clark and Summers (1990), in which an individual's behavior according

to their individual characteristics can be represented by a transition probability matrix, or Markov matrix. Thus, it portrays the dynamic mobility of individuals from occupations with different levels of qualification - high, medium and low - by the probability of

SLIDE 5

5

individual i being in state k in period t + 1, conditional on the fact that he/she is in state l in period t. The transitions between states are treated as a Markov process in which a steady state is reached, regardless of initial conditions and the steady state ratio of each state is found as a function of the entire transition matrix. In addition, a simulation exercise is performed to measure how the transition probabilities of a category can influence the fractions of time in each state of the other. The second methodology is based on a multivariate econometric analysis, using a Multinomial Logit model. With this model, it is possible to understand how the set of individual and macroeconomic variables affect the probability of an individual to be in

• ccupations of higher or lower qualification. To do so, occupations are classified in socio-
• ccupational status groups, based on the classification proposed in Monsueto (2016),

which considers aspects of productivity and quality of jobs. The classes are given by: High/Medium-High, Medium, Medium-Low/Low. This classification can be used to assess, for example, the possibilities of occupational mobility and escape from low- income situations. Thus, analyzing the occupations that are used as a gateway to the labor market can help to anticipate the type of occupational trajectory that these individuals may experience. Results Starting in mid-2003, the labor market responded positively to the expansion of the Brazilian economy, generating a near-continuous trend of falling unemployment even shortly after the financial crisis of 2008. At the same time, the period also experienced an increase in the number of hiring’s through the signed work portfolio, contributing to a significant reduction in informality. On the other hand, despite the homogenization

• bserved at the end of the investigation period, there is a clear difference in levels of

unemployment according to gender, showing a clear advantage for white male workers. In general, black women continue to experience the greatest difficulties in the metropolitan Brazil labor market, even when unemployment rates are controlled by the years of schooling.

SLIDE 6

6

Figure 1 - Unemployment rate, by educational groups, Metropolitan Brazil, 2003-2014

Source: PME, 2003-2014

When analyzing the inactivity rates, the main differences remain according to gender, with greater similarity between white and non-white individuals within each sex. The fall in unemployment in the period seems to have helped to maintain stable rates of inactivity for the working age population, and the greatest fluctuations must have occurred precisely in the transition between the occupations with different types of qualification. This result shows the importance of analyzing in more detail how both the transition of groups between different occupations occurs, as well as the determinants of entry into one of these situations. Figure 2 - Inactivity rate by educational groups, Metropolitan Brazil, 2003-2014

Source: PME, 2003-2014

10 20 30 40 50 60 2002 2005 2008 2011 2014

0 a 4 years of study

White men Non-white men White women Non-white women 10 20 30 40 50 60 2002 2005 2008 2011 2014

5 a 7 years of study

White men Non-white men White women Non-white women 10 20 30 40 50 60 2002 2005 2008 2011 2014

8 a 10 years of study

White men Non-white men White women Non-white women 10 20 30 40 50 60 2002 2005 2008 2011 2014

11 or more years of study

White men Non-white men White women Non-white women

SLIDE 7

7

Figure 3 shows the result of the Markov transition matrices. The estimates of the percentage of average time spent in each group of occupations by qualification (high, medium and low), disaggregated by color and gender, can see below. Whites stand out for their greater distribution of time in occupations with high qualification. Figure 3 - Fraction of time estimated for each occupational class, Metropolitan Brazil, 2003-2016

Fonte: PME, 2003-2016

Assigning the probability pattern of occupational transition from whites to nonwhites shows that the fraction of time in which the latter pass in the best occupations increases, while the time spent in the two categories of occupations of lower socioeconomic quality is reduced. This result seems to indicate that if both whites and nonwhites had the same

• pportunities or patterns of occupational transition, both groups could use mobility as a

source of exit from situations of greater precariousness or low income at the same

• intensity. However, what is seen in practice is that nonwhite workers of both sexes spend

more time in activities of lower socioeconomic quality, resulting in an increase in the level of occupational segregation in the country.

SLIDE 8

8

Figure 4 - Variation in the fraction time in each occupational class for non-whites to the probability of occupation of white, Metropolitan Brazil, 2003-2016

Fonte: PME, 2003-2016

Finally, multinomial logit models are estimated for the probability of observing an

• ccupied worker acting in one of the three major categories of occupational quality, as a

function of a series of control variables. Among these factors, the greatest interest is to analyze the role of the occupation in which he was in the previous period, looking for evidence of temporal dependence. The estimated coefficients are displayed in Tables 1 and 2, separated by gender and color groups. The estimated models show a significant temporal dependence of the contemporary

• ccupational situation of the workers, evidencing difficulties to carry out transitions

between the segments of socioeconomic quality considered. However, when comparing the size of the coefficients that indicate the position occupied in the previous period, it can be seen that nonwhite workers have a greater temporal dependence on the past situation, making it difficult to enter occupations of better status and the lower quality. An active employment policy that generates new vacancies and opportunities only, or priority, in occupations of this nature can even be understood as short-term solutions, removing individuals from situations of unemployment and extreme poverty, since they are the cheapest ones for immediate employment. But in the long run they only tend to reproduce the dynamics of low labor productivity, leading to a perpetuation of social inequality between whites and nonwhites and increased occupational segregation.

SLIDE 9

9

Table 1 - Coefficients estimated by the multinomial logit model - High occupation

Non-white women White women Non-white men White men Medium occup. in t

• 47.668*
• 50.654*
• 45.812*
• 48.615*

(0.02) (0.03) (0.02) (0.03) Low occup. in t

• 44.621*
• 47.921*
• 42.631*
• 45.715*

(0.05) (0.05) (0.04) (0.06) 0/3 years of study 0.456*

• 2.054*

0.422*

• 2.343*

(0.07) (0.12) (0.06) (0.09) 4/7 years of study 0.289*

• 2.062*

0.290*

• 2.429*

(0.05) (0.07) (0.05) (0.05) 8/10 years of study 0.280*

• 1.970*

0.205*

• 2.312*

(0.06) (0.07) (0.05) (0.05) 11 or more years of study 0.297*

• 1.938*
• 0.068
• 2.228*

(0.05) (0.04) (0.05) (0.04) Spouse

• 0.048

0.019 0.059

• 0.015

(0.03) (0.04) (0.04) (0.05) Son 0.168*

• 0.293*

0.285*

• 0.218*

(0.05) (0.06) (0.04) (0.05) Another parent 0.222*

• 0.167

0.273*

• 0.238*

(0.08) (0.11) (0.08) (0.09) 35/44 years old

• 0.337*

0.065

• 0.376*

0.080*** (0.04) (0.05) (0.03) (0.04) 45/54 years old

• 0.545*

0.073

• 0.649*

0.098** (0.04) (0.05) (0.03) (0.04) 55/65 years old

• 0.673*

0.097***

• 0.872*

0.244* (0.04) (0.05) (0.04) (0.05) RMRE 0.029

• 0.103
• 0.174*

0.015 (0.06) (0.09) (0.05) (0.08) RMSA

• 0.039

0.052

• 0.116*

0.085 (0.05) (0.10) (0.04) (0.09) RMBH

• 0.139*
• 0.028
• 0.328*

0.124* (0.04) (0.05) (0.04) (0.05) RMRJ

• 0.112*
• 0.007
• 0.267*
• 0.028

(0.04) (0.05) (0.03) (0.04) RMPOA 0.032

• 0.157*

0.176**

• 0.124*

(0.07) (0.05) (0.08) (0.04) a2004 0.026 0.084 0.207*

• 0.004

(0.09) (0.10) (0.08) (0.09) a2005 0.227*

• 0.003

0.191* 0.002 (0.09) (0.09) (0.07) (0.09) a2006 0.191**

• 0.096

0.263*

• 0.089

(0.09) (0.09) (0.07) (0.09) a2007 0.220** 0.049 0.329*

• 0.102

(0.09) (0.09) (0.07) (0.09) a2008 0.324* 0.038 0.365*

• 0.104

(0.09) (0.09) (0.07) (0.09) a2009 0.173*** 0.042 0.342*

• 0.171***

(0.09) (0.09) (0.07) (0.09) a2010 0.341*

• 0.175**

0.346*

• 0.148***

(0.09) (0.09) (0.07) (0.09) a2011 0.232** 0.073 0.489*

• 0.135

SLIDE 10

10 (0.09) (0.09) (0.07) (0.09) a2012 0.194**

• 0.033

0.501*

• 0.170***

(0.10) (0.09) (0.07) (0.09) a2013 0.309*

• 0.033

0.348*

• 0.153***

(0.09) (0.09) (0.08) (0.09) a2014 0.360* 0.111 0.471*

• 0.092

(0.09) (0.09) (0.07) (0.08) a2015 0.298* 0.025 0.464*

• 0.157***

(0.09) (0.09) (0.07) (0.09) a2016 0.229

• 0.021

0.340**

• 0.050

(0.20) (0.21) (0.14) (0.17) Intercept 24.553* 25.594* 23.569* 24.928* (0.10) (0.09) (0.08) (0.09) *** p<0.10, ** p<0.05, * p<0.01

Table 2 - Coefficients estimated by the multinomial logit model – Low occupation

Non-white women White women Non-white men White men Medium occup. in t

• 3.456*
• 3.606*
• 3.617*
• 3.871*

(0.04) (0.04) (0.04) (0.04) Low occup. in t 3.624* 3.785* 3.912* 4.204* (0.04) (0.04) (0.04) (0.04) 0/3 years of study 1.068* 0.284 0.827* 0.303*** (0.21) (0.21) (0.19) (0.17) 4/7 years of study 0.904* 0.481* 0.522* 0.177 (0.15) (0.12) (0.15) (0.13) 8/10 years of study 0.676* 0.562* 0.481* 0.265** (0.15) (0.12) (0.15) (0.13) 11 or more years of study 0.576* 0.349* 0.376* 0.168 (0.13) (0.10) (0.14) (0.12) Spouse

• 0.022

0.011

• 0.096
• 0.016

(0.08) (0.08) (0.10) (0.12) Son

• 0.000

0.060

• 0.079

0.044 (0.12) (0.13) (0.10) (0.12) Another parent 0.111 0.103 0.028

• 0.370***

(0.21) (0.22) (0.20) (0.22) 35/44 years old

• 0.154
• 0.258*

0.030

• 0.131

(0.10) (0.10) (0.09) (0.10) 45/54 years old

• 0.504*
• 0.450*
• 0.015
• 0.205**

(0.10) (0.10) (0.09) (0.10) 55/65 years old

• 0.670*
• 0.600*
• 0.137
• 0.228***

(0.13) (0.11) (0.11) (0.12) RMRE 0.048

• 0.153

0.144 0.029 (0.15) (0.20) (0.12) (0.21) RMSA 0.085 0.015 0.237**

• 0.348

(0.12) (0.21) (0.11) (0.28) RMBH 0.032

• 0.226***

0.076

• 0.122

(0.11) (0.12) (0.09) (0.12) RMRJ

• 0.183***
• 0.136
• 0.055
• 0.231**

(0.11) (0.10) (0.09) (0.10)

SLIDE 11

11 RMPOA 0.121

• 0.093
• 0.077
• 0.107

(0.19) (0.10) (0.18) (0.09) a2004 0.039 0.402*** 0.084

• 0.033

(0.25) (0.23) (0.20) (0.23) a2005 0.265 0.278 0.288

• 0.081

(0.25) (0.22) (0.19) (0.21) a2006 0.111 0.192 0.348***

• 0.117

(0.24) (0.22) (0.19) (0.21) a2007 0.323 0.544* 0.273 0.064 (0.24) (0.21) (0.18) (0.21) a2008 0.372 0.542* 0.220

• 0.049

(0.25) (0.21) (0.19) (0.20) a2009 0.045 0.575* 0.185

• 0.132

(0.25) (0.22) (0.19) (0.21) a2010 0.225 0.139 0.087

• 0.085

(0.25) (0.22) (0.19) (0.20) a2011 0.172 0.600* 0.063

• 0.142

(0.24) (0.22) (0.19) (0.22) a2012 0.068 0.379*** 0.094

• 0.162

(0.25) (0.21) (0.19) (0.22) a2013 0.331 0.325 0.115

• 0.125

(0.25) (0.23) (0.20) (0.22) a2014 0.474** 0.745* 0.272

• 0.016

(0.24) (0.21) (0.18) (0.20) a2015 0.316 0.504** 0.408**

• 0.184

(0.23) (0.22) (0.18) (0.21) a2016 0.206 0.296 0.215 0.274 (0.49) (0.49) (0.36) (0.34) Intercept

• 0.782*
• 0.694*
• 1.077*
• 0.332

(0.26) (0.22) (0.22) (0.22) Pseudo - R2 0.9001 0.9338 0.8982 0.9418 Number of obs. 33329 51849 48596 66702 Chi2 11610743.75 23397659.12 10084155.82 27120647.52 Prob>Chi2 0.00 0.00 0.00 0.00 *** p<0.10, ** p<0.05, * p<0.01

Final Considerations This article intends to study occupational transitions with different levels of qualification, by gender and color, in the period between 2002 and 2016. From the data of the Monthly Employment Survey applied to six metropolitan regions of the country, the results show that whites individuals stand out for their greater distribution of time in occupations with high qualification. Assigning the probability pattern of occupational transition from whites to nonwhites shows that the fraction of time in which the latter pass in the best occupations increases,

SLIDE 12

12

while the time spent in the two categories of occupations of lower socioeconomic quality is reduced. In addition, the multinomial logit models show a significant temporal dependence of the contemporary occupational situation of the workers, evidencing difficulties to carry out transitions between the segments of socioeconomic quality

• considered. This result is more expressive to non-white workers.

In this way, the results presented in this article show evidence that the country can see an increase in the levels of occupational segregation in the next years, given the greater difficulty of certain demographic groups to ascending jobs with better socioeconomic

• quality. Therefore, there is the real need to promote public policies of better occupational

insertion, promoting the entry into occupations of social profile and more stable professional path. References AMADEO, E.; CAMARGO, J. M.; GONZAGA, G. BARROS, R. E MENDONÇA, R. A natureza e o funcionamento do mercado de trabalho brasileiro. Texto para Discussão,

• 353. IPEA, Rio de Janeiro, 1994.

ARAÚJO, V. F., RIBEIRO, E. P. Diferenciais de salários por gênero no Brasil: uma análise regional. Textos para Discussão n.2001/11. UFRGS – PPGE, Porto Alegre, 2001. BARROS, P. B., MENDONÇA, R. S. P. Os determinantes da desigualdade no Brasil. Texto para Discussão, 377. IPEA, , Rio de Janeiro, 1995. BARROS, P. B.; MACHADO, A.F.; MENDONÇA, R.S.P. A desigualdade da pobreza: estratégias ocupacionais e diferenciais por gênero, Texto para Discussão, 453. IPEA, Rio de Janeiro, 1997. CLARK, Kim B.; SUMMERS, Lawrence H. Unemployment insurance and labor market

• transitions. In: SUMMERS, L. H. Understanding unemployment. Cambridge: MIT, 1990.

CONTINI, B., VILLOSIO, C. Job change and wage dynamics. Working Paper Series nº5, Laboratorio R. Revelli Centre for Employment Studies, 2000. DAVIA, M.A. Studying the impact of job mobility on wage growth at the beginning of the employment career in Spain. In: IX Encuentro de Economía Aplicada, Jaén, 2006.

• FITZENBERGER. B., KUNZE, A. Vocational training and gender: wages and
• ccupational mobility among young workers. Discussion Paper No. 05-66, ZEW, 2005.

GABRIEL, P.E. An examination of occupational mobility among full-time workers, Monthly Labor Review, Bureau of Labor Statistics, 129(9), 32-36, 2003. GARCIA PEREZ, J.I., SANZ, Y. R. Wages changes through job mobility in Europe: a multinomial endogenous switching approach. Labour Economics, Elsevier, 12(2005), p.531-555, 2005.

SLIDE 13

13

HOLZER, H.J.; LANE, J.I.; VILHUBER, L. Escaping low earnings: the role of employer characteristics and changes. Discussion Paper nº 1269-03, Institute of Research on Poverty, 2003. IBGE. Série histórica da pesquisa mensal de emprego. Disponível em <www.ibge.gov.br> MALTSEVA, I. Gender differences in occupational mobility and segregation at the labor market: the case of Russian economy. Working Paper Series n. 05/11, Education and Research Consortium, 2005. MATOS, R. S. E MACHADO, A. F. Diferencial de rendimento por cor e sexo no Brasil (1987-2001). Econômica, Universidade Federal Fluminense, 8(1), p.5-27, 2006. McLAUGHLIN, K.J. A theory of quits and layoffs with efficient turnover, Journal of Political Economy, University of Chicago Press, 99(1), 1-29, 1991. MONSUETO, Sandro Eduardo. NOTA TÉCNICA EM ECONOMIA n. 07 NERI, M.; COELHO, D.; ANCORA, M. y PINTO, A. Aspectos dinâmicos do desemprego e da posição na ocupação. In: Encontro Nacional de Estudos do Trabalho, ABET, Rio de Janeiro, 1997. OLIVEIRA, A. M. H. C. Indicadores da segregação ocupacional por sexo no Brasil. In: XI Encontro Nacional de Estudos Populacionais da ABEP, Associação Brasileira de Estudos Populacionais, 1998. OLIVEIRA, A. M. H. C. A segregação ocupacional por gênero e seus efeitos sobre os salários no Brasil. In: WAJNMAN, S. e MACHADO, A.F. (Orgs.). Mercado de trabalho – uma análise a partir das pesquisas domiciliares no Brasil, Editora UFMG, Belo Horizonte, pp.121-149, 2003. OLIVEIRA, A. M. H. C., RIBEIRO, P. M. Diferenciais ocupacionais por raça e gênero no mercado de trabalho metropolitano no Brasil. In: XI Encontro Nacional de Estudos Populacionais da ABEP, Associação Brasileira de Estudos Populacionais, 1998. PACI, P., SERNEELS, P (Ed.). Employment and shared growth: rethinking the role of labor mobility for development. Washington: The World Bank, 2007. PARRADO, E.; CANER, A. e WOLF, E. Occupational and industrial mobility in the United States, Labour Economics, Elsevier, 14(3), 435-455, 2007. PARRADO, E. e WOLFF, E. Occupational and industry mobility in the United States, 1969-1992, C.V. Starr Center for Applied Economics, New York University, Working Paper 99-20. 1999. PINTO, A., NERI, M. Mobilidade ocupacional e raça: origens, destinos e riscos dos afro-

• brasileiros. Série Ensaios Econômicos nº 392, EPGE/FGV, Rio de Janeiro, 2000.

RAMOS, L., VIEIRA, M. L. Desigualdade de rendimentos no Brasil nas décadas de 80 e 90: evolução e principais determinantes. Texto para Discussão, 803, IPEA, Rio de Janeiro, 2001. SOARES, C., OLIVEIRA, S. Gênero, estrutura ocupacional e diferenciais de rendimento. Econômica, Universidade Federal Fluminense, 6(1), p.5-33, 2004.