As “heranças do passado”

1
As “heranças do passado”:
mudanças estruturais e mobilidade social de filhos de agricultores no Brasil (1973-1996).
Carlos Antonio Costa Ribeiro (UERJ)
Artigo a ser apresentado no XXVI encontro da ANPOCS, Caxambú, MG.
GT-07: Estratificação e Mobilidade Social
2
1 – Introduction:
In this paper, I will describe structural effects on social mobility from 1973 to
1996. Such structural effects are better observed via analyses of absolute mobility rates
and “structural mobility” patterns which are both conditioned by the relative size of
classes in origin and destination distributions. As I have argued in the third chapter, the
most striking characteristic of the Brazilian mobility data is the disparity between origin
and destination distributions. Due to steady industrialization and urbanization in Brazil
since the 1950’s the distribution of class origins is extremely different from the
distribution of class destination. In particular, the majority of workers has origins in rural
classes and destinations in manual and non-manual urban class positions. This
dissimilarity determines high levels of absolute mobility rates. One way of describing
these high levels of mobility is to unveil “structural mobility” forces, which provoke a
great amount of mobility out of the rural classes into the urban classes. There are still
other ways of describing absolute mobility rates which illuminate other aspects of the
class structure. Outflows rates, for example, describe the chances of mobility for men
with origins in each class to move to the various destination classes.
The analyses in this chapter will allow me to further develop the descriptions
initiated in the previous chapter. In other words, I will examine in what measure changes
in the Brazilian class structure directly affect patterns of social mobility. As I have been
arguing, two aspects are of major interest. First, I will investigate the impact of the fast
transition from an agricultural and rural to an industrial and urban society on the levels
and rates of absolute mobility. Second, I must determine if the context of fast economic
growth until 1980 and economic stagnation afterwards can be also observed in patterns of
3
social mobility. While in the previous chapter I described the demographic, economic,
and geographical aspects of the Brazilian transition, as well as changes in the origin and
destination class distributions, in the present chapter I will depict the patterns of social
mobility related to these broader changes. The study of absolute mobility rates will unveil
the temporal aspect of the Brazilian class structure. What are the class destinations of
men coming from different class origins? Which are the class origins of men in each class
in 1973, 1982, 1988, and 1996? Which are the effects of the disparity between origin and
destination class distributions on social mobility rates? These are the questions I answer
in this chapter.
In the first section, I will describe the levels of upward and downward mobility in
1973, 1982, 1988 and 1996. There is a great interest by readers of a mobility study in
knowing if vertical mobility increased or decreased. However, as it should be clear at this
point of this dissertation, a mobility study can give much more information than simple
description of upward and downward mobility. In particular, a mobility study is a good
way of describing the class structure of a given society and the degree of openness of this
class structure. This does not mean, of course, that the determination of levels of upward
and downward mobility is not relevant, but only that it is just one part, and perhaps not
the most important one, of what is been pursued herein.
In the second part of the chapter, I will describe mobility chances of men with
origins in the eleven CASMIN classes. These mobility chances will be described using
the outflow rates (row percentages in the mobility table). Outflow rates are an excellent
and simple way of observing the chances of people coming from diverse origins of
reaching different class destinations, but they are not a measurement of the degree of
4
openness of a society (this is the topic of the next chapter). Despite the fact that these
chances, measured by outflows, are influenced by the relative size of destination and
origin classes and thus are not shaped to describe degrees of openness, they are
interesting because they indicate the actual amount of mobility experienced by people
coming from diverse origins.
Finally, in the third part of the chapter, I will analyze the consequences of the
disparity between origins and destinations for the levels of mobility observed. For doing
so, I will describe the supply and demand of class positions and the patterns of “structural
mobility” as defined by special log-linear models.
2 – Total and Vertical Mobility
The proportion of men who were found in a different class position than their
fathers, the total mobility rate, suffered a significant increase from 1973 to 1982 in
Brazil. According to the eleven classes schema, in 1973 about 62 percent of all men were
found in a different class position than their fathers, while after 1982 this figure increased
to 69 percent. These indices demonstrate that class societies are really different from
caste societies. Class is not fixed at birth and people have chances of moving up or down
in the class structure.
Nevertheless, for a good deal of the movement observed in the eleven class
schema it is not easy to characterize it as “upward” or “downward.” As I have extensively
discussed in the second chapter, the hierarchical order of classes within the CASMIN
schema is not unambiguous. Therefore, the determination of the direction of observed
mobility must be examined via some type of aggregation of classes, that is, the movement
5
among some classes cannot be considered strictly as upward or downward. Following the
description in the second chapter and the representation in figure 2.1, I suggest the
following hierarchical levels from upper to lower classes: (1st) I; (2nd) II + IVa; (3rd) IIIa
+ IVb + V; (4th) IIIb + VI + IVc1; and (5th) VIIa + VIIb. According to this classification,
mobility among certain classes, for example between class II and IVa or VIIa and VIIb,
cannot be defined as vertical mobility, but only as an horizontal movement within the
class structure. In contrast, mobility from classes among the five levels defined above
shall be defined as vertical mobility either upward or downward.
In table 4.1, I decompose the total mobility indices obtained from the eleven
classes into total vertical and non-vertical (according to the five levels defined in the
previous paragraph). And, then, I further decompose total vertical mobility into total
upward and downward1. This simple exercise provides some interesting indications about
Table 4.1
Decomposition of total mobility rates (TMR) into total vertical (TV) and
total non-vertical (TNV) mobility and of total vertical mobility into
total upward (TU) and total downward (TD) mobility.
Survey Year TMR
TV
TNV TV/TNV
TU
TD
TU/TD
1973
1982
1988
1996
62
69
68
69
43
49
50
51
19
20
18
18
2.3
2.5
2.7
2.9
33
38
39
38
10
11
11
13
3.3
3.5
3.5
2.9
possible trends in upward and downward mobility from 1973 to 1996 in Brazil.
Considering the first stage of this decomposition, as shown in the second, third
and fourth columns of table 4.1, it is striking to see that there is no change in the level of
vertical mobility and non-vertical mobility from 1973 to 1996. In other words, although
1
See Marshal, Swift and Roberts 1997 for an exposition of this methodology.
6
the total level of mobility increased across the years, this increase is characterized by
constancy of non-vertical and vertical mobility. In particular, the fourth column, showing
the rate of total vertical by total non-vertical mobility (TV/TNV), suggests that there is
practically no change in total vertical mobility when compared to total non-vertical
mobility.
It is in the second stage of the decomposition, as shown in columns fifth, sixth,
and seventh, that one can indeed observe if vertical mobility is upward or downward in
its character. The data proves that upward mobility is at least three times more likely than
downward mobility in each one of the four years studied, and that the chances of
downward mobility vis-à-vis upward mobility do not change over the years. Indeed, in
1973 there were three times more chances of experiencing upward rather than downward
mobility, and in 1996 this rate continued to indicate that upward mobility was three times
more likely than downward mobility.
In conclusion, in spite of the increase in total mobility, upward mobility continued
to be more frequent than downward mobility and vertical mobility remained constant
across the years. The types of decomposition described in table 4.1 indicate, in fact, a
striking tendency for constancy in patterns of vertical mobility from 1973 to 1996. It is
impressive that even during a period of steady economic growth, industrialization, and
urbanization the levels of vertical mobility remained constant in Brazil.
3 - The Extent of Social Mobility:
How do mobility and immobility chances vary by origin class and historical
periods? In order to answer this question I will examine outflow rates (row percentages in
7
the mobility table) of men with origins in three broadly defined types of work, and with
origins in eleven classes. The answer may vary according to the two levels of aggregation
that will be examined. This apparent paradox is simply a result of the level of detail used
to describe mobility and immobility chances. Both types of description are valid for as
each provides different kind of information. While the division between non-manual
(sum of classes: I, II, IIIa, IVa, IVb, and IIIb), urban manual (sum of classes: V, VI, and
VIIa), and farm work (class IVc plus VIIb) reveals almost constant mobility chances in
the period of 1973 to 1996; the analyses of the mobility chances of men with origins in
eleven different classes indicates fluctuations in this period.
Before presenting these data, I briefly explain how mobility and immobility
chances will be analyzed in this section. My instrument is outflow rates, row percentages
in the mobility table measuring the mobility and immobility chances of men coming from
different origins. Outflow rates allow a clear and easy observation of the chances of
mobility for people coming from each class origin, but they do not provide any relative
measurement. That is, outflows show which are the observable amount of mobility and
immobility for each one of the class origins, but do not indicate which are the relative
mobility chances of people with different origins. These relative chances, which are
crucial to measure both inequality of opportunity and the association between origins and
destinations, will be described in the following chapter. Although technically simpler
than relative chances, outflows give precise figures describing the proportion of people
coming from each origin class that reached each one of the destination classes.
Using outflow rates in this section I will draw conclusions about changes and
continuities in mobility chances in the periods of fast structural changes up to 1982 and
8
of economic stagnation afterwards. These mobility chances were affected by the fast
industrialization that occurred in the post-war period in Brazil. I begin the description of
mobility and immobility chances by presenting the mobility profile of men from nonmanual, urban manual and farm origins in the period of 1973 to 1996. Table 4.2
summarizes this data.
Table 4.2
Outflow: Origin Type of Work by Survey Year by Current Type of Work:
Men Aged 20-64, Brazil, 1973, 1982, 1988, and 1996.
Origin Type of Work
Current Type of Work
Year
Non-manual
Manual
Farm
Total
Non-Manual Work
1973
0.68
0.28
0.05
1.00
1982
0.67
0.29
0.04
1.00
1988
0.68
0.28
0.04
1.00
1996
0.66
0.31
0.04
1.00
Urban Manual Work
1973
1982
1988
1996
0.32
0.34
0.36
0.34
0.62
0.62
0.59
0.62
0.06
0.04
0.04
0.04
1.00
1.00
1.00
1.00
Farm Work
1973
1982
1988
1996
0.13
0.18
0.19
0.19
0.38
0.45
0.44
0.44
0.49
0.37
0.37
0.37
1.00
1.00
1.00
1.00
Total
1973
1982
1988
1996
0.23
0.28
0.31
0.31
0.39
0.46
0.45
0.47
0.38
0.26
0.24
0.22
1.00
1.00
1.00
1.00
The most striking characteristic of this table is the constancy in mobility and
immobility chances across the four years analyzed. The only exception is the change in
mobility chances of sons of farm workers from 1973 to 1982. In 1973, only 13 percent of
the sons of farm workers reached non-manual positions, in 1982 these chances increased
to 18 percent. The chances of reaching urban manual occupations also increased from 38
percent in 1973 to 45 percent in 1982. Finally, the immobility chances of sons of farm
9
workers decreased from 49 percent in 1973 to 37 percent in 1982. These change in the
mobility profile of farm workers sons are certainly determined by the sharp decrease in
the rural sector of the economy that occurred in the 1970’s. After 1982, there are no
changes in the mobility chances of farm workers’ sons.
All other mobility chances observable in table 4.2 display a remarkable constancy
across the years. From 1973 to 1996, only one out of three sons of manual workers are
mobile to non-manual occupations; and about two out of three are immobile in manual
occupations - the number moving to farm positions is minimal. In this same period, two
out of three sons of non-manual workers inherit non-manual working position and one
out of three move to urban manual working jobs. Although this stability sounds
remarkable, it is essential to examine a more detailed mobility/immobility picture.
In the remaining of this section, I will describe the mobility chances of men
coming from eleven different origin classes. Men coming from certain origin classes will
have some particular economic, cultural and social resources in their background. The
description of the general characteristics of each one of the eleven classes will help to
understand the possible influence that the resources associated to each class could have
on mobility chances that are displayed in table 4.3.
10
Table 4.3
Outflow: Class of Origin by Survey Year by Current Class: men aged 20-64, Brazil, 1973, 1982, 1988, and 1996.
Current Class
Origin
Code Year
I
II
IIIa
IVa
IVb
IIIb
V
VI
VIIa IVc1
Higher-grade Profes. & Adm.
I 1973 40.0 17.8 15.2 4.3
2.2
3.3
4.1
3.7
7.4
1.3
1982 40.3 12.6 13.6 8.4
2.1
3.8
4.9
4.7
6.3
1.7
1988 35.9 12.3 13.4 11.6 2.5
4.2
4.8
4.8
6.7
1.8
1996 37.6 8.7 14.2 11.8 3.6
6.0
3.5
4.4
7.6
1.3
Lower-grade Profes. & Adm.
VIIb
0.7
1.7
2.1
1.3
Total
100.0
100.0
100.0
100.0
II 1973
1982
1988
1996
22.7
22.2
21.1
19.5
23.1
18.4
18.5
14.9
18.1
17.7
17.6
14.4
2.9
8.7
9.8
9.7
3.7
2.3
3.9
6.8
4.2
5.8
5.4
5.7
4.2
5.3
5.1
2.9
11.6
7.7
6.8
12.1
8.3
9.7
8.4
11.9
0.4
1.3
1.3
0.4
0.8
0.9
2.0
1.8
100.0
100.0
100.0
100.0
Higher-grade Routine non-manual
IIIa 1973
1982
1988
1996
17.2
14.5
12.9
13.0
18.6
13.3
12.5
10.2
23.2
22.1
23.8
20.3
2.8
7.4
6.3
6.6
6.0
2.2
4.2
5.1
4.6
6.0
7.6
10.1
3.4
6.8
5.0
4.1
9.4
12.0
13.0
12.0
12.6
13.1
12.5
15.6
0.6
0.7
0.7
0.6
1.6
2.0
1.6
2.5
100.0
100.0
100.0
100.0
"Small" proprietors, with employees
IVa 1973
1982
1988
1996
11.8
14.9
14.8
11.8
12.1
11.7
12.3
8.2
11.6
14.4
10.2
10.1
16.8
20.8
26.3
19.4
8.5
4.0
5.0
8.9
3.7
4.5
4.4
6.8
1.9
3.3
4.1
3.1
13.1
10.5
7.9
12.7
14.1
12.0
10.7
14.9
1.3
1.2
1.7
1.7
5.2
2.6
2.5
2.4
100.0
100.0
100.0
100.0
"Small" proprietors, without employees
IVb 1973 12.3
1982 11.1
1988 10.9
1996 8.8
11.0
10.8
9.9
7.2
14.2
12.0
12.1
11.3
4.4
11.0
10.1
13.2
15.8
11.2
13.1
14.2
5.2
5.9
5.5
7.5
3.4
3.8
4.5
2.8
10.0
13.3
13.0
11.6
17.1
16.4
15.5
17.5
2.5
1.1
1.3
1.4
4.1
3.5
4.0
4.4
100.0
100.0
100.0
100.0
Lower-grade Routine non-manual
IIIb 1973 11.9
1982 11.0
1988 9.6
1996 8.6
14.8
10.5
10.2
9.0
15.2
16.4
15.5
12.2
1.8
5.8
8.3
5.5
7.2
2.9
5.7
8.2
5.1
10.1
10.8
12.8
4.0
4.6
4.5
3.6
19.1
15.3
14.3
16.6
18.8
20.9
18.1
20.5
0.3
0.6
0.4
2.2
2.2
2.4
2.5
100.0
100.0
100.0
100.0
Technicians and super. manual work
V 1973
1982
1988
1996
8.2
7.7
9.4
7.1
12.1
9.9
11.0
7.5
17.0
14.2
12.0
12.1
2.6
6.4
5.7
7.1
4.4
1.8
3.7
6.5
3.9
6.7
7.7
8.1
8.5
9.5
8.3
6.3
21.9
21.3
21.3
23.7
19.3
18.9
16.9
18.7
0.3
0.6
0.2
0.6
2.1
2.9
3.7
2.2
100.0
100.0
100.0
100.0
Skilled manual workers
VI 1973
1982
1988
1996
3.4
3.9
3.8
3.3
8.0
6.5
6.9
5.4
10.7
10.5
9.5
7.2
2.0
4.8
5.5
4.9
3.5
2.7
3.8
4.4
3.9
6.0
5.9
7.5
4.6
4.5
4.9
3.0
36.3
36.2
33.6
36.0
21.9
21.3
22.1
24.4
0.8
0.3
0.4
0.2
4.9
3.3
3.6
3.6
100.0
100.0
100.0
100.0
Semi- & unskilled manual workers
VIIa 1973
1982
1988
1996
2.7
3.2
3.1
3.0
6.4
5.7
7.0
6.2
8.9
9.6
10.1
8.0
1.3
4.0
4.8
4.0
4.6
3.4
3.8
4.6
4.8
6.2
6.8
8.0
4.3
4.4
4.3
2.7
24.7
25.2
21.5
22.7
36.2
34.8
33.9
36.4
0.7
0.3
0.5
0.3
5.3
3.3
4.2
4.0
100.0
100.0
100.0
100.0
Rural employers
IVc1 1973
1982
1988
1996
4.6
5.2
8.5
4.1
4.2
4.9
5.7
3.5
6.3
5.7
6.0
6.6
1.9
6.9
8.6
7.2
7.1
5.9
4.6
7.1
2.4
3.7
4.3
4.5
2.0
2.2
3.6
2.0
10.6
12.8
12.5
15.2
16.7
19.7
17.2
23.5
23.6
10.4
13.7
8.0
20.6
22.5
15.3
18.5
100.0
100.0
100.0
100.0
Agricultural workers
VIIb 1973
1982
1988
1996
0.6
1.0
1.3
1.2
1.2
1.7
2.1
2.1
2.2
3.2
2.9
2.9
1.0
2.3
3.1
2.9
4.6
4.2
5.4
5.5
1.9
3.4
3.2
3.8
1.5
1.7
2.1
1.4
15.2
19.4
17.9
17.8
23.6
25.5
24.6
25.0
4.8
2.1
2.1
2.2
43.4
35.5
35.4
35.1
100.0
100.0
100.0
100.0
1973
1982
1988
1996
3.2
4.1
4.3
4.0
4.0
4.4
5.0
4.3
5.4
6.5
6.5
6.1
1.7
4.4
5.2
5.0
5.3
4.3
5.2
5.9
2.7
4.3
4.4
5.6
2.3
2.7
3.2
2.2
16.8
20.0
18.8
19.9
22.9
23.7
23.1
24.6
5.0
2.2
1.9
1.7
30.8
23.4
22.3
20.8
100.0
100.0
100.0
100.0
Total
11
First, I direct my attention to the higher professional and managerial group (class
I). This occupational group includes all higher-grade professionals, self-employed or
salaried, and higher-grade public service administrators. The employment status of higher
service class workers in the private sector tends to be ambiguous as in the case of
company directors who are stockholders, or self-employed professionals (physicians,
engineering consultants and etc). Occupations within this class often offer their
incumbents incomes that are high, secure and likely to increase over their lifetimes.
These positions typically require the exercise of authority and offer considerable
autonomy and freedom from control by others. This class together with class II is in
general conceived as the “service class” of modern capitalist society - that is, the class
position whose members exercise power and expertise on behalf of corporate bodies. It
also includes independent businessmen and professionals. Members of this class are in
general able to provide material, cultural, and social support to their offspring. The
chances of immobility and/or short distance downward mobility are typically high.
Indeed, although the level of immobility in class I decreases from 40 percent in
1973 to 37 percent in 1996, it continues to be the non-rural class with the highest degree
of immobility. In addition, the degree of mobility to the lower professional and
managerial group (class II) was around 18 percent in 1973 and declined gradually to 8.7
percent in 1996. In other words, the total outflow rate of men with origin in the highergrade professional and administrative occupations (class I) to the service class as a whole
(I+II) declined gradually from 58 percent in 1973 to 46 percent in 1996. This decrease is
compensated by a gradual increase from 4 percent in 1973 to 11.8 percent in 1996 in the
proportion found in the ranks of the employer petite bourgeoisie (class IVa). As a
12
consequence, across the four samples around 60 percent of the workers whose fathers’
were higher-grade professionals and administrators remain in the service class (I+II) or
move to urban employer positions (class IVa). Indeed, the proportion of men whose
fathers were higher-grade professionals or administrators that remain in white-collar
occupations (classes I, II, IIIa, IIIb, IVa, and IVb) stays stable from 1973 to 1996 in about
80 percent. In each one of the samples, long rage downward mobility to the class of
unskilled manual workers (classes VIIa and VIIb) occurs only for about 8 percent of the
men with origins in class I. In sum, sons of higher-grade professionals and administrators
display a remarkable stability in their mobility chances (outflow patterns) between 1973
and 1996. The main characteristic of this stability is the fact that the majority of men with
origins in class I tend to stay in the highest occupational categories, and to avoid long
range downward mobility. It is relevant to highlight, however, that there is a tendency
over time for men coming from class I to move to occupations in the group of proprietors
with employees (class IVa), rather than to inherit service class occupations (I+II). This
mobility into the ranks of proprietors with employees can only be considered as short
distance downward mobility, because this type of proprietors has high standards of living.
Lower professional and managerial positions (class II) include lower rank
professionals; higher-grade technicians; lower grade administrators, officials and
managers; and supervisors of non-manual employees. In general, these positions
guarantee income that is just below the previous group and employment benefits such as
pension entitlements. In bureaucratic hierarchies, these occupations stand in middle and
lower ranges and imply the exercise of some authority and discretion at the same time
that they are typically subjected to supervision. The pattern of mobility of the upper (I)
13
and lower (II) professional and administrators group differ in a number of important
aspects. For example, the degree of immobility in the lower professional and
administrative group declined faster than in the upper group. In 1973, among sons of
lower grade professional and administrators 23 percent remained in this same class
position, while in 1996 only 15 percent followed in their fathers’ footsteps. The
percentage found either in the lower or the higher professional and administrative classes
(I+II) declined from 46 percent in 1973 to 34 percent in 1996. This decline is in part
compensated by an increase in the percentage of men going to petite bourgeoisie
positions (classes IVa and IVb), that is, in 1973 only 6.6 percent moved to petite
bourgeoisie positions, while in 1996 about 16 percent moved to these positions. The
percentage remaining in non-manual positions (classes I, II, IIIa, IIIb, IVa, and IVb)
continued around 75 percent from 1973 to 1988 and declined to 71 percent in 1996. The
percentage going to skilled manual positions (classes V and VI) remained around 15
percent over the samples. Finally, the percentage experiencing long rage downward
mobility to the class of unskilled manual workers (classes VIIa and VIIb) increased from
9 percent in 1973 to 14 percent in 1996. In sum, sons of lower-grade professionals and
administrators tend to remain less frequently in the highest class position then sons of
higher-grade professionals and administrators. Although there was an increase over time
in the proportion experiencing long range downward mobility into non-skilled manual
classes, the majority of mobile men with origins in the lower service class (II) still tend to
be more likely to stay in higher than in lower hierarchical positions in the class structure.
The two routine non-manual classes (IIIa and IIIb) have different positions in the
class structure. Since the activities and duties of higher-grade routine non-manual
14
workers (IIIa) are of primary importance in supporting the activity of “service class
people” (classes I and II), this type of routine non-manual workers usually enjoy many
privileges in their working conditions. Class IIIa comprises occupations such as clerical
workers, cashiers, commercial travelers, etc. In contrast, lower-grade routine non-manual
workers (class IIIb) are in the lowest hierarchical positions in the non-manual work
group, and could be considered as a non-manual working class position in the sense that
their labor contract is typically characterized by a very specific exchange of wages for
efforts. Class IIIb comprises occupations such as hairdressers, shop salesmen and
assistants, telephone operators, barmen, etc.
The fact that the higher routine non-manual group (IIIa) appears to occupy a
position in the hierarchy just below the professional and managerial classes is reflected in
their mobility chances as displayed in table 4.3. In 1973, about 36 per cent of those from
higher routine non-manual origin were found in the professional and managerial groups
(I+II). Although this percentage declined to 23 percent in 1996, men with origins in class
IIIa continued to move more often to professional and managerial positions than men
coming from any other non-managerial or non-professional origin. Another important
characteristic is the high percentage of immobility within non-manual occupations
(I+II+IIIaIVa+IVb+IIIb). In 1973, around 72 percent with origins in class IIIa remained
in non-manual occupations, this figure declined gradually to 65 percent in 1996. After
higher and lower grade professionals and administrators, the group of higher-grade
routine non-manual workers (together with the group of small proprietors with
employees) is the one with highest rates of mobility to non-manual occupations. In
addition, the patterns of outflow to manual occupations in 1973 are similar to those of
15
sons of lower grade professionals and administrators, that is, about 25 percent of the men
coming from each one of these origins moved down to manual occupations. From 1982
onward, the proportion of men whose fathers were higher-grade routine non-manual
workers that moved down to manual occupations increased to 32 percent, but the
proportion of men with origins in the lower service class moving down to manual
working positions continued around 25 percent. Although there are some differences
between the patterns of outflow of men whose fathers were lower service class workers
(class II) and higher routine non-manual workers (class IIIa), I would say that outflow
patterns of these two groups have more in common then outflow patterns of the two
routine non-manual groups.
Perhaps the best way to contrast workers with higher-grade and lower-grade
routine non-manual origins is to observe the rates of outflow to manual occupations
(V+VI+VIIa+VIIb) and to professional and managerial positions (I+II). Over the years
the proportion of workers with upper routine non-manual origins moving to manual
occupations (V+VI+VIIa+VIIb) increased from 25 to 32 percent. In contrast, the
proportion of those with origin in the lower routine non-manual group moving to manual
occupations remained around 45 percent from 1973 to 1996. Although outflow rates into
professional and managerial occupations declined over time for both higher and lowergrade routine non-manual, they continued to be more frequent for men coming from the
higher-grade group. In 1973, one out of three of those from higher-grade routine nonmanual origin (IIIa) and only one out of four with lower-grade routine non-manual
origins (IIIb) moved to professional and managerial positions (I+II). In 1996, one out of
four coming from class IIIa and less than one out of five coming from class IIIb moved to
16
service class positions. In other words, men with origins in the higher routine non-manual
class (IIIa) are more likely to move up to the top of the class hierarchy and less likely to
move down to manual classes than are men with origins in lower routine non-manual
occupations (IIIb).
The “small” proprietors or petit bourgeois categories (classes IVa and IVb)
include proprietors (except rural proprietors), self-employed artisans and many other
“own account workers” apart from professionals. Class IVa also includes medium
proprietors because the Brazilian data does not allow a distinction between medium and
small business. The market situation of these workers is distinctive because of their
employment status, although income levels display considerable variability. Economic
security is in general less predictable than in the case of salaried employees. The upper
stratum is composed by employers and stands in a high position in the hierarchy of
classes. By contrast, the lower stratum is composed by self-employed “small” proprietors
without formal employees.
Despite the fact that these two positions enjoy on average different standards of
living, men from employer and self-employed petite-bourgeoisie origins (IVa and IVb)
display very similar outflow rates. However, there seems to be a small difference in trend
directions. From 1973 to 1988, the proportion of sons of proprietors with employee
staying in non-manual occupations increased from 64 to 73 percent and in 1996 this
proportion diminished to 65 percent. In other words, in 1982 and 1988 sons of proprietors
with employees had slightly better chances to avoid downward mobility to manual
occupations then they had in the other years. In contrast, the proportion of sons of self-
17
employed proprietors moving to non-manual occupations remained around 62 percent
across the samples.
Workers in the class position of lower grade technicians (class V), whose work is
from a manual character, tend to earn high income, to have some degree of formal
education, and to enjoy reasonable security of employment. Technicians and supervisors
of manual work are certainly the elite of the manual working class. This status can be
observed in the outflow rates of sons of technicians and supervisors. Among the manual
working classes, sons of technicians and supervisors have the highest outflow rates into
professional and managerial occupations (I+II). Outflow into the service class stays
around 20 percent from 1973 to 1988, and declined to 14 percent in 1996. Although this
outflow declines after 1988, it remains at least twice as large as the outflow of any other
manual group into the service class. Conversely, sons of technicians and supervisors have
the lowest rates among manual workers of outflow into other manual position. Over the
samples at least 65 percent of the workers with origins in skilled or unskilled urban
manual classes (VI and VIIa) remain in manual working and 80 percent or more of those
men with origins in the rural manual working class (VIIb) remain in manual occupations.
In contrast, only 50 percent of those whose fathers were technicians or supervisors of
manual work remain in manual working positions.
The distinction between skilled and unskilled urban manual workers seems to be
more tenuous in Brazil than in other countries. This fact is due to the very low rates of
formal or vocational training that skilled workers have in Brazil. The class of skilled
manual workers (class VI) comprises all those industrial activities requiring some form of
manual skills that could be obtained even through informal apprenticeship. Most skilled
18
manual workers are in “modern industrial sectors” (steel, oil, automobile, and etc),
although some of them are also in the construction sector. By contrast, unskilled manual
workers are typically responsible for all sorts of manual occupations that do not require
special qualification. Although they are concentrated in “traditional industries”, they are
also found in many other industrial sectors, as well as in the lower services sector of the
economy. Both skilled and unskilled manual employees typically work under very strict
labor contracts characterized by a direct exchange of wages for tasks. Despite all
differences between skilled and unskilled manual working positions, they have more
characteristics in common between themselves, than with the class of lower grade
technicians (V).
Indeed, outflow patterns of men with skilled manual origins (class VI) are closer
to the patterns of men with origins in unskilled manual work (class VIIa) than to the
patterns of sons of technicians and supervisors of manual work (class V). Over the
samples, the proportions of sons of skilled (VI) and unskilled (VIIa) manual workers
found in service position (I+II) stays around 9-10 percent, in routine non-manual
positions (IIIa+IIIb) around 15-16 percent, and in petit bourgeoisie positions (IVa+IVb)
around 6-9 percent. Moreover, 35 percent of the sons of skilled and unskilled manual
workers (VI and VIIa) move up to non-manual occupations (I+II+IIIa+IIIb+IVa+IVb),
and around 70 percent remains in manual occupations (V+VI+VIIa+VIIb). Immobility
for sons of skilled manual workers and unskilled manual workers remains respectively
around 35 percent, that is, about half from the 70 percent remaining in manual working is
immobile. In sum, unskilled and skilled manual occupations have not only very similar
outflow patterns, but also these patterns are very stable from 1973 to 1996.
19
Finally, as previously suggested, farmers and rural workers display a distinctive
mobility profile. While class IVc includes all small farmers who have employees, class
VIIb includes all manual rural employees and small farmers without employees
producing mainly for subsistence. The first characteristic in mobility pattern of these
classes that has to be noticed, and that I already mentioned in the beginning of this
section, is that from 1973 to 1982 there was a drop in the proportion of workers with
origins in the rural sector, and that the figures for 1988 and 1996 are very similar to those
from 1982. In 1973, about 50 percent of the sons of rural workers and 44 percent of the
sons of rural employers remained in the rural sector. From 1982 onward, the number of
men with origins in the rural employer class that remained in the rural sector declined
slowly from 33 percent in 1982 to 26 percent in 1996. Over the samples the majority (90
percent or more) of the sons of rural workers remaining in the rural sector continued to be
found in the ranks of rural working occupations (class VIIb). In contrast, more than half
of the sons of rural employers who remained pursuing farming activities moved to rural
working positions. This apparent downward mobility of sons of rural employers to rural
working positions is probably a reflection of the fact that many sons work in their fathers’
farms. What is remarkable is that the mobility from rural working origins into rural
employer destinations is practically non-existent. In 1973, only one out of ten sons of
rural workers who stayed in rural sector occupations moved up to rural employer
positions, and in 1982 and after, only one of each seventeen sons of rural workers staying
in the rural sector moved up to rural employer positions.
There are also some marked differences in the outflow rates of sons of farmers
and of rural workers that moved out of the rural sector. About half of the sons of rural
20
employers (class IVc) that moved to urban occupations entered non-manual occupations
– these proportions do not change from 1973 to 1996. In contrast, among those sons of
rural workers (class VIIb) who went to urban occupations only one-fifth in 1973 and onefourth in the subsequent years achieved non-manual positions. Indeed, there is a
remarkable stability in the low chances of sons of rural workers (class VIIb) to move out
of manual occupations: in 1973 only 17 percent and in 1996 only 20 percent moved out
of manual working positions. Moreover, from those sons of rural workers staying in
manual occupations only 2 out of 10 in 1973 and 1.5 out of 10 after 1982 moved up to
qualified manual working positions (classes V and VI).
I began this section asking: how mobility and immobility chances vary by origin
class and historical period. A short answer for this question is simply that mobility
chances vary a lot among classes and not so much overtime, the more elaborated and
convincing answer was sought in the exhaustive descriptions rendered in this section. The
summary of the detailed and lengthy answer is the following.
It is clear that mobility within the broad divisions of manual, non-manual and
farm working do not vary over the years, with the exception of a significant decline of
immobility in farm working from 1973 to 1982. By contrast, the description of mobility
patterns for men coming from eleven different origin classes indicates that there is a
considerable amount and variability in mobility in the period described, and that most of
the movement is in the upward direction, although this upward mobility is of a short
distance character. In other words, men move but move in general to the classes just
above them in the hierarchical structure of classes. In addition, after 1982 there seems to
be a tendency for increasing downward mobility for classes I, II, IIIa, and IIIb, which is
21
compensated by an increase in upward mobility for classes IVa and IVb. Finally, there is
a remarkable stability in the immobility rates of manual workers and an alarming
tendency for farm workers to stay in rural or urban manual working positions.
Although stability is the outstanding characteristics in mobility chances in the
period of 1973 to 1996, there is yet a tendency of downward mobility for white-collar
workers and of upward mobility for proprietors after 1982. However, it is difficult to
determine if these tendencies are a consequence of the period of economic stagnation
after 1982 or if they are simple “trend-less” fluctuations. It is also difficult to determine
whether the decline of immobility in the rural sector until 1982 and its stability
afterwards is a consequence of the period of steady industrialization until 1982 and
economic stagnation in the 1980’s and 1990’s. The consequences of the decline in
opportunities in the rural sector for social mobility shall be examined in the next section.
4 - Disparities between Origin and Destination Class Distributions, and Social Mobility:
One of the distinguishing features of the Brazilian mobility data in 1973, 1982,
1988, and 1996 is the disparity between the origin and the destination class distributions.
Broadly speaking, this disparity is characterized by a surplus of farmers’ and farm
workers’ sons, and a surplus of vacancies in urban manual working and white-collar
positions in the destination distribution. This disparity, which is a direct consequence of
the swift and relatively recent industrialization in Brazil, impels a great amount of social
mobility. In this section, I will answer questions about the impact of industrialization (by
elaborating upon the disparity between origin and destination class distributions) on
22
patterns of social mobility in Brazil. I will examine both patterns of class supply and
recruitment, and patterns of “structural mobility.”
4.1 - Supply and Recruitment
The patterns of class supply and recruitment are related not only to the topics of
immobility and mobility (both represented by outflow rates), but also to the rates of “selfrecruitment” (represented by inflow rates), and the dissimilarity between origin and
destination distributions for each class. Immobility is defined by the proportion of men
from a given class origin who stay in these same class as adults. In other words,
immobility is the percentage outflow into the same class destination as the origin. “Selfrecruitment” is defined by the proportion of men in a given destination class who come
from this same class origin, or in technical terms, by the percentage inflow (column
percentage) in the same class. Finally, the index of dissimilarity between origin and
destination distributions for each class is the sum of positive differences between outflow
(row percentages) and inflow (column percentage) for this class. If the index of
dissimilarity for a given class is low, then this class recruits from the same classes to
which it sends their sons. Conversely, if the index of dissimilarity for a given class is
high, then this class has divergent patterns of supply and recruitment. That is, it supplies
sons for certain classes and recruits from certain other classes.
Table 4.4 presents “immobility” and “self-recruitment” rates, as well as the index
of dissimilarity for each class in 1973, 1982, 1988, and 1996. According to this table,
each one of these three indices varies more across classes than across survey years. It is
probably the case that patterns of self-recruitment, immobility, and dissimilarity for each
23
class do not differ across the years.2 This constancy suggests that the impact of the
disparity between origins and destinations class distributions on mobility rates does not
change in the period of 1973 to 1996. Hence, one could reasonably say that the impact of
the transition from an agrarian into an industrialism society on patterns of class supply
and recruitment continued to be significant across the period of 1973 to 1996. My
description of supply and recruitment using the indices in table 4.4 will disregard the
possible small differences among the four years, and will concentrate on the differences
among classes.
Table 4.4
Self-recruitment, Immobility, and Index of Dissimilarity between Origin and Destination for Class j (j); by Origin and Survey Year:
men aged 20-64, Brazil, 1973, 1982, 1988, and 1996.
Self-recruitment
Immobility
j
(inflow)
(outflow)
Origin
Year
Year
Year
1973 1982 1988 1996 1973 1982 1988 1996 1973 1982 1988 1996
Higher-grade Profes. & Adm.
I 15.1 14.7 13.9 16.1
40.0 40.3 35.9 37.6
42.5 39.6 34.7 32.1
Lower-grade Profes. & Adm.
II 7.2
6.4
7.5
6.6
23.1 18.4 18.5 14.9
45.6 46.9 46.1 37.9
Higher-grade Routine non-manual
IIIa 7.7
6.6
8.8
8.5
23.2 22.1 23.8 20.3
46.4 43.2 39.4 34.5
"Small" proprietors, with employees
IVa 13.8 15.2 14.8 11.6
16.8 20.8 26.3 19.4
44.2 46.3 43.3 36.6
"Small" proprietors, without employees IVb 15.2 12.7 12.1 11.4
15.8 11.2 13.1 14.2
58.6 63.9 56.6 48.1
Lower-grade Routine non-manual
IIIb 1.4
3.4
3.7
5.1
5.1 10.1 10.8 12.8
53.4 53.4 42.4 37.2
Technicians and super. manual work
Skilled manual workers
Semi- & unskilled manual workers
Rural employers
Agricultural workers
Overall
V
VI
VIIa
IVc1
VIIb
3.7
17.5
12.1
31.8
91.7
3.8
18.2
12.8
33.6
88.8
3.9
20.0
16.3
20.8
91.9
4.6
24.0
18.4
15.8
90.0
8.5
36.3
36.2
23.6
43.4
9.5
36.2
34.8
10.4
35.5
8.3
33.6
33.9
13.7
35.4
6.3
36.0
36.4
8.0
35.1
49.9
57.2
66.1
49.9
48.3
42.4
59.5
65.4
57.2
58.1
40.5
53.0
58.9
55.5
56.6
34.6
46.4
52.8
58.7
55.6
65.0
58.6
57.8
53.2
38.0
31.0
31.0
31.0
36.0
40.7
36.6
34.1
It is widely recognized in the scholarly literature on social mobility that some
classes serve as channels for upward mobility (i.e., Hout, 1989; Blau and Duncan, 1967;
Glass, 1964; Bogda, and Wesolowski, 1986; and Heath, 1981). These classes usually
have a low proportion of self-recruitment and a low proportion of immobility. A
considerable portion of men with origins in these classes tend to move up to more
privileged class positions, while most men recruited to these classes come from lower
2
Indeed, the correlation coefficient of self-recruitment, immobility, and indices of dissimilarity
distributions is always bigger than .82 for each two years combination. The only exception is the
24
classes. Consequently, these classes have divergent patterns of supply and recruitment,
meaning that their index of dissimilarity is typically low in comparison to other classes.
As can be easily grasped from table 4.4, in Brazil there are at least two classes that
clearly function as channels for upward mobility: technicians and supervisors of manual
work (class V) and lower-grade routine non-manual (class IIIb). Both recruit more than
half of their members from lower class origins (VI, VIIa, and VIIb) and send at least 25
percent or more of their sons’ to classes on the top of the hierarchy (I, II, and IIIa) (Table
A4.1 in the appendix to this chapter presents inflow rates for the eleven classes). It is
noteworthy, that the two classes function as channels to upward mobility for they are
precisely at threshold from manual to non-manual work. While class IIIb can be
considered as the less privileged of the non-manual working positions, class V is
undoubtedly the elite of the manual working class. It is easier, thus, to sons of manual
workers and farm workers to achieve positions as lower-grade technician (V) or lowergrade routine non-manual worker (IIIb), than to make long distance mobility travels to
petite bourgeoisie (IVa+b) or upper white-collar (I, II, and IIIa) positions. In turn, sons of
lower-grade technicians and lower routine non-manual workers have better chances of
moving all the way up in the classes hierarchy, than other manual and farm workers do.
Table 4.4 also indicates that some classes of origin display high levels of “selfrecruitment” taking in few sons from other origins and, at the same time, sending large
proportions of sons to other classes. Agricultural workers (VIIb) and to a lesser extend
farm employers (IVc1) are such type of classes. For example, the majority of sons of
agricultural workers are outside agriculture (57 percent or more of mobility over the
correlation between indices of dissimilarity between 1973 and 1996 (r = .51), but even in this case there is
correlation.
25
samples), but nearly every agricultural worker has origins in this same class (around 90
percent are self-recruited over the samples). This pattern means that although agricultural
workers send their offspring to many other class positions, they are isolated in the
agricultural sector because they do not recruit from other class positions. One shall
remember that, according to the outflow rates described in the previous section of this
chapter, the majority of the sons of agricultural workers are mobile to urban manual
working class. However, looking from the perspective of recruitment patterns of all
classes, the limited mobility chances of agricultural workers’ sons could be thought of as
a paradox. All classes, except for class I, recruit at least 20 percent of their members from
the ranks of agricultural working class (VIIb). Rather than a paradox, this incongruity
between the patterns of mobility (outflows) and recruitment (inflow) related to
agricultural workers result from the fact that a vast amount of workers have origins in
class VIIb (65 percent in 1973 and 53.2 percent in 1996) and the number of positions in
agricultural work in the contemporary class distribution is limited (30 percent in 1973
and 20 percent after 1982). In other words, patterns of recruitment in the Brazilian data
are marked by the disparity between the high proportion of men with origins in
agricultural classes and the relatively low proportion of men with destination in these
classes. As it will be seen further ahead in this chapter, this disparity is properly
described as “structural mobility” forces favoring mobility out of rural classes and into all
other classes.
In contrast to the depleted rural working class, urban manual working classes
expanded very quickly over the years. According to the figures in table 4.4, the urban
manual working classes (VI and VIIa) are characterized by high immobility and self-
26
recruitment, and by divergent patterns of supply and recruitment. In fact, these classes
recruit most of their members from farm working positions (IVc or VIIb), and keep more
than half of their offspring. This pattern indicates that it is easy for sons of farm workers
to enter into urban manual positions, and difficult to sons of manual workers to leave the
ranks of the manual working class.
Another consequence of rapid industrialization for the patterns of supply and
recruitment in Brazil is that petite bourgeoisie (IVa and IVb) and upper white-collar (I, II,
and IIIa) classes do not have high rates of self-recruitment. Since these classes expanded
very quickly in the post-war period, they cannot rely only on self-recruitment and need to
recruit their members from other origins. Even the class of higher-grade professionals
and managers (class I) - that could be seen as the elite in the class schema I am using - is
characterized by a heterogeneous pattern of recruitment. Indeed, the fact that only around
15 percent of higher-grade professionals and managers (class I) are self-recruited attests
against the so-called elite closure thesis in the Brazilian case (i.e., Parkin, 1979; and
Weber, 1978a). According to this thesis, most elite members come from their own ranks.
It is relevant to remember, however, that in Brazil downward mobility for class I is
characterized by short distance to other white-collar positions, rather than by long
distance movements to manual working class. That is, class I recruits from many
different classes but retain most of its sons. The relatively low degree of self-recruitment
in Brazil when compared to other countries should be interpreted as a consequence of the
recent transition into an industrial society that helped to expand rapidly the size of class I
(see Costa-Ribeiro and Scalon, 2001). The same conclusion applies to the supply and
27
recruitment patterns of the other white-collar (II and IIIa) and proprietorial classes (IVa
and IVb).
The patterns of class supply and recruitment described in this section lead to three
general conclusions concerning the impact of the transition into industrialism on social
mobility patterns in Brazil. First, the rapid industrialization and urbanization in the postwar period continued having an impact in the imbalance between class supply and
recruitment at least until 1996. Second, this imbalance favors upward mobility. Third, the
fast transition created more opportunities in white and blue-collar classes than the supply
coming from these same classes. This imbalance necessitated heterogeneous patterns of
recruitment in non-rural classes, to the extent that, a great proportion of the workers in
the contemporary class distribution, even in the highest positions, come from lower class
origins and from rural class origins.
4.2 - Structural Mobility
The best way to describe the impact of the disparity between the contemporary
distribution of classes and the distribution of the origin classes on mobility patterns is the
measurement of “structural mobility”. As a direct consequence of this disparity, structural
mobility creates opportunities in some destinations and limits them in other destinations,
regardless of social origins. Therefore, in this section I describe the social mobility that is
determined by these structural forces. This type of mobility is usually high in societies
going through major transitions like the one Brazil experienced since the 1950’s with
steady industrialization and urbanization. Previous studies on Brazilian mobility patterns
(Hutchinson, 1958, and 1960; Pastore, 1981; Scalon, 1999; and Pastore and Silva, 2000)
28
have concluded that there are high levels of upward mobility, which are due rather to
structural changes than to an increasing openness of the class structure. In this section
and the next chapter, I not only confirm this previous expectation, but also explain in
more detail what are the consequences of structural change and degrees of openness for
the Brazilian mobility regime.
In order to analyze “structural mobility” in this section, I will draw on the
framework developed by Sobel, Hout, and Duncan (1985). According to these authors,
structural mobility redistributes workers of fixed origins into the destination class
distribution independently of social origins. The analyses of this type of mobility depend
on the fitting of the log-linear model of quasi-symmetry with or without restrictions to
block asymmetric association to the mobility table.
Although the quasi-symmetry models applied to each one of the four Brazilian
mobility tables account for more than 70 percent of the association under the baseline
model of quasi-perfect mobility, they still do not fit the data (see table A4.2 in the
appendix). However, according to the BIC criterion3, a quasi-symmetry model fit the data
if two asymmetric associations between class I and II, and class I and V are considered.
For this last model of quasi-symmetry with constrains for two asymmetric associations,
the fits are the following: in 1973, L2 = 216.9, df = 48, BIC = -289.4, and Index of
Dissimilarity () = 3.7; in 1982, L2 = 515.3, df = 48, BIC = -15, and  = 5.5; in 1988, L2
= 224.3, df = 48, BIC = -276.5, and  = 5.4; and in 1996, L2 = 255, df = 48, BIC = 253.5, and  = 6.2. In addition, these quasi-symmetry models with constrains for two
3
A negative number for BIC favors the acceptance of the model instead of the saturated model.
29
asymmetries account in each one of the years for more than 93 percent of the association
under the baseline model of quasi-independence.
In order to directly compare the four tables for 1973, 1982, 1988, and 1996, I used
the model of Conditional Quasi-symmetry with homogeneous association (CQSHA)
proposed by Sobel (1988). This model states not only that the association between origin
and destination is symmetric, but also that it is constant across the four mobility tables. It
is not applied to each table at a time, but rather to the four tables at the same time, that is,
the model is fitted to the three-way table cross-classifying 11 classes of origin by 11
classes of destination by 4 survey years. The CQSHA model does not fit the four
Brazilian mobility tables (L2 = 7289.8, df = 350, BIC = 3065, and  = 3.4). However, if
the two asymmetric associations between class I and II, and class I and V suggested
above are included in each one of the four origin by destination sub-tables the model fits
the 11x11x4 table according to the BIC criterion (L2 = 1785.1, df = 348, BIC = -2,416,
and  = 2.5). In addition, the CQSHA model with constrains for two asymmetries
accounts for 97 percent of the association under the baseline model of independence (see
table 5.4 in the next chapter). It is important to observe that due to the fact that the
Brazilian mobility tables have a large N (174,880 when considering the four tables) it is
extremely difficult to find a model that fits the data according to any 2 statistics because
this type of goodness-of-fit measurement is very sensitive to large sample sizes.
Nevertheless, it would be precipitated to accept the fit of the CQSHA model
before testing if the strength of the association changes over time. In order to test the
hypothesis that the strength of association changes across the four mobility tables, I
applied Xie’s (1992) log-multiplicative model specifying the quasi-symmetry association
30
with two asymmetric terms just discussed. According to hierarchical comparison, the logmultiplicative model improves the fit over the CQSHA with two asymmetries model (see
respectively models M3x and M3 in table 5.4 of the next chapter): L2 = [1,785.00 –
1631.12] = 153.88 with d.f. = 4. In addition, the log-multiplicative form explains 97.3
percent of the association not accounted for by the Independence model and has a more
negative BIC (-2,522) than the CQSHA model (BIC = -2,416). As a consequence, I
selected this log-multiplicative form of the quasi-symmetry model with two asymmetries
to describe the patterns of structural mobility in this chapter. In the next chapter, I will
describe patterns of association. It is important to note that the strength of association is
declining over time according to the findings of the log-multiplicative model.
As a consequence of the asymmetric associations, the description of structural
mobility will not consider structural mobility into classes I, II, and V. Nevertheless,
structural mobility ratios [j/i] for cells in other columns for classes IIIa, IVa, IVb, IIIb,
VI, VIIa, IVc, and VIIb are meaningful. The log of structural mobility ratios [log(j/i)]
measures the effect of structural mobility (departure from symmetric association due to
marginal disparity) on flows between pairs of class category (Hout 1989: 86; and Luijkx,
1994).
Table 4.5 presents structural mobility log-ratios and indicates in which cells there
is a significant inflow or an outflow due to structural mobility. The cells containing a
number are the ones presenting an inflow caused by structural mobility, and the empty
cells are the ones in which there is an outflow due to structural mobility. The magnitude
of an inflow represented in the cells containing a number bellow the main diagonal are
equivalent to the magnitude of an outflow in the corresponding cell above the main
31
diagonal. For example, in 1973, the number in the fourth row, third column [cell (4,3)] is
1.27 and corresponds to an inflow into destination class IIIa (column 3) from origin class
IVa (row 4) with a magnitude of 1.27. Conversely, the equivalent cell (3,4)
corresponding to row 3, column 4 in the other side of the diagonal has no value on it. The
suppressed value is -1.27 and corresponds to an outflow from class IIIa (row 3) to class
IVa with value -1.27. The meaning of these values is the anti-log of the numbers
considered. For instance, in 1973 the expected mobility from class IVa into class IIIa
[cell (4,3)] is (e1.27 =) 3.6 times greater than the expected mobility in the opposite
direction from class IIIa into class IVa.
32
Table 4.5
Estimates of structural mobility [log(j/i)] from the log-multiplicative model quasi-symmetry with constraints on
(21 = 1/12) and (71 = 1/17) (model M3x in table 5.4) by survey years 1973, 1982, 1988, and 1996: men aged 20-64, Brazil.
Current Class (j)
Origin (i)
Code
I
II
IIIa
IVa
IVb
IIIb
V
VI
VIIa IVc1 VIIb
1973
Higher-grade Profes. & Adm.
I
Lower-grade Profes. & Adm.
II ...
Higher-grade Routine non-manual
IIIa ...
...
"Small" proprietors, with employees
IVa ...
...
1.27
1.07 0.72 0.13 0.33
"Small" proprietors, without employees
IVb ...
...
1.95 0.68
1.75 1.40 0.81 1.01
Lower-grade Routine non-manual
IIIb ...
...
0.21
Technicians and super. manual work
Skilled manual workers
Semi- & unskilled manual workers
V
VI
VIIa
...
...
...
...
...
...
0.55
1.14
0.94
0.35
0.93
0.73
...
...
Rural employers
Agricultural workers
IVc1
VIIb
...
...
...
...
3.45
4.21
2.17 1.493 3.24
2.94 2.26 4.01
...
...
1982
Higher-grade Profes. & Adm.
Lower-grade Profes. & Adm.
Higher-grade Routine non-manual
"Small" proprietors, with employees
"Small" proprietors, without employees
Lower-grade Routine non-manual
I
II
IIIa
IVa
IVb
IIIb
...
...
...
...
...
...
...
...
...
1.22
2.21
0.51
1.00
Technicians and super. manual work
Skilled manual workers
Semi- & unskilled manual workers
V
VI
VIIa
...
...
...
...
...
...
0.56
1.28
1.15
0.06
Rural employers
Agricultural workers
IVc1
VIIb
...
...
...
...
4.65
4.63
3.44
3.41
1988
Higher-grade Profes. & Adm.
Lower-grade Profes. & Adm.
Higher-grade Routine non-manual
"Small" proprietors, with employees
"Small" proprietors, without employees
Lower-grade Routine non-manual
I
II
IIIa
IVa
IVb
IIIb
...
...
...
...
...
...
...
...
...
0.63
1.70
0.31
1.06
Technicians and super. manual work
Skilled manual workers
Semi- & unskilled manual workers
V
VI
VIIa
...
...
...
...
...
...
0.52
1.28
1.23
0.64
0.59
Rural employers
Agricultural workers
IVc1
VIIb
...
...
...
...
3.43
4.43
2.79
3.80
1996
Higher-grade Profes. & Adm.
Lower-grade Profes. & Adm.
Higher-grade Routine non-manual
"Small" proprietors, with employees
"Small" proprietors, without employees
Lower-grade Routine non-manual
I
II
IIIa
IVa
IVb
IIIb
...
...
...
...
...
...
...
...
...
0.53
1.32
0.27
0.79
Technicians and super. manual work
Skilled manual workers
Semi- & unskilled manual workers
V
VI
VIIa
...
...
...
...
...
...
0.79
1.15
1.00
0.26
0.62
0.47
Rural employers
Agricultural workers
IVc1
VIIb
...
...
...
...
3.41
4.05
2.89
3.52
-
0.2
-
2.304 2.50
3.07 3.27
0.71
1.70
-
0.66
1.65
0.048
0.77
0.64
...
...
-
0.132
-
2.44
2.42
4.14
4.12
...
...
3.37
3.35
3.50
3.48
-
0.33
1.39
-
0.11
1.17
0.42
0.47
0.22
0.97
0.92
...
...
-
0.05
-
1.73
2.73
3.12
4.13
...
...
2.15
3.15
2.20
3.20
-
0.26
1.06
-
0.53
0.17
0.32
0.53
0.88
0.74
...
...
-
0.15
-
3.15
3.78
...
...
2.27
2.90
2.41
3.05
-
2.09
2.73
0.93
0.77
-
0.07
1.06
-
0.02
-
1.00
-
0.63
-
33
According to the magnitude of the estimated log-ratios in table 4.5 it is possible to
support the hypothesis that there are no great differences in terms of structural mobility
among these four years. I assert that the forces of structural mobility are similar in 1973,
1982, 1988, and 1996. In other words, the impact of the fast transition from a mainly
agricultural into an industrial society apparently continues to have effects on the Brazilian
mobility patterns from 1973 to 1996. Or, in other words, the disparity between origin and
destination distribution is the main factor determining high volumes of observed mobility
in Brazil. If this is indeed the case, what are the consequences one should draw from the
estimated log-ratios in table 4.5?
The first information that I must highlight is that log-ratios’ values in the two last
rows (excluding the two last columns) in each year stand out by their high values. These
high values - ranging from 1.49 in cell (10,5) in 1973 to 4.65 in cell (10, 3) in 1982 indicate that structural mobility forces (disparity between origin and destination class
distributions) impel high volumes of mobility out of the two rural classes (IVc and VIIb)
into classes IIIa, IVa, IVb, IIIb, VI, and VIIa. Or, in other words, the high proportions of
outflow from the rural classes are equivalent to high inflows from of rural workers in all
other classes. The log-ratios ranging from 1.49 to 4.65 for mobility out of the rural
classes indicate that structural forces compel men to move out of rural classes into the
other manual and non-manual classes at a rate ranging from 4 to 105 times greater than
the back mobility flow from those other classes into rural classes. For example, log-ratios
indicate that the chances of farmers’ (IVc1) and farm workers’ (VIIb) sons to move into
higher-grade routine non-manual positions (IIIa) are at least 30 times greater than the
mobility chances in the reverse order (from IIIa into rural classes). And the chances of
34
men with origins in rural classes to move into urban manual working classes (VI and
VIIa) is at least 8 times greater than mobility in the reverse order (from VI and VIIa into
rural classes). The massive mobility out of rural classes should be expected since the
transition from an agricultural and rural society into an industrial and urban one was
amazingly fast and relatively recent in Brazil. Consequently, the diminishment of
working opportunities in the rural sector was compensated by the creation of
opportunities in white-collar and urban manual working.
The log-ratio for manual-to-manual mobility (classes VI, and VIIa) ranging from
.05 to .20 across the years show that structural forces favor inflows into unskilled manual
working class positions (VIIa) from skilled working class positions (VI). As I have
indicated in the previous sections, these two manual working classes seem to have very
similar mobility profiles and the structural forces favoring mobility into class VIIa cannot
be strictly defined as a force favoring downward mobility within manual working
positions. In the case of the higher routine non-manual class (IIIa), log-ratios ranging
from .21 to 2.21 show that structural forces are favoring mobility into this class for men
coming from classes IVa, IVb, IIIb, V, VI, and VIIa rather than in the other direction by
at most 23 percent (e.21 =1.23). Structural forces also favor mobility from lower-grade
petite bourgeoisie (class IVb) into manual working positions (V, VI, and VIIa) rather
than mobility in the other direction, that is, from urban manual working classes into
lower-grade petite bourgeoisie.
In the most general terms, structural mobility favors upward mobility flows. The
high volume of mobility out of the rural classes into other classes is certainly the most
appealing conclusion that can be draw from table 4.5. However, it is essential to be clear
35
that the high volumes of mobility out of the rural sector into white-collar and manual
working classes is determined by structural forces diminishing opportunities in the
primary sector and expanding opportunities in other sectors rather than by an increasing
openness of the stratification system. The question of social openness will begin to be
tackled in the next chapter.
The analyses of “structural mobility” show that the high levels of observed
mobility and of upward mobility described in this chapter are a direct consequence of
structural forces or the disparity between origins and destination class distributions.
Although the analyses in the next chapter indicate that some of the mobility is probably
due to a decrease in association between origin and destination classes, the findings based
on the analyses of structural mobility lead to the conclusion that almost the totality of the
high levels of absolute mobility observed is a direct consequence of the dissimilarity
between origin and destination classes distributions. This dissimilarity is such that large
amounts of workers are impelled out of the rural class origins into all other class
positions. However, one must bear in mind that the majority of workers are still in
manual working classes throughout the period of 1973 to 1996.
36
5 – Summary and conclusion:
In this chapter I investigated the patterns of absolute or gross mobility in Brazil
from 1973 to 1996. Therefore, this chapter complements the descriptions of the Brazilian
class structure pursued in the last chapter simply because absolute mobility rates are
useful to determine not only the extent of social mobility, but also the class background
of workers in the contemporary class distribution. Absolute mobility is always influenced
by the relative sizes of each class in origin and destination. Or, in other words, disparities
between origin and destination class distributions have a direct and clear effect on
absolute mobility rates. Since this disparity, in Brazil, results from the steady
industrialization and urbanization that occurred from the 1950 to 1980, and to a lesser
extent from the economic crisis after 1980, my investigation was designed to unveil the
possible effects of these historical factors on gross or absolute mobility patterns. In order
to pursue these objectives my strategy was: (1) to define variation on the degree of
upward and downward mobility over time, (2) to describe in detail the extent of gross or
absolute mobility using outflow rates, and (3) to determine in what measure the patterns
of gross mobility are provoked by the dissimilarity between origin and destination class
distributions. The overall conclusion is that the transition from an agricultural into an
industrial society continued to have significant impacts on absolute mobility patterns
throughout the period from 1973 to 1996. As for the impacts of the period of economic
expansion until 1980 and the crisis afterwards, it was not possible to observe any
significant differences in absolute mobility rates.
(1) Although the total mobility rate in Brazil increased from 62 to 69 percent
between 1973 and 1982, the analyses of vertical mobility indicated that there was no
37
change in the degree of upward mobility. Since upward mobility remain three times more
likely than downward mobility across the years from 1973 to 1996, I conclude that
neither economic growth until 1982, nor economic crisis after 1982 had an impact on the
levels of vertical mobility. The increase in total mobility from 1973 to 1982 must be
attributed to a decrease in the level of immobility in the rural classes. This interpretation
is confirmed by the facts that immobility in manual and non-manual types of work did
not change from 1973 to 1996, and that mobility into rural work from non-rural work is
almost inexistent.
(2) The description of mobility patterns (outflow rates) of men coming from
eleven different origin classes shows that there is a considerable variability in mobility
during the period described, and that short distance upward mobility prevails. After 1982,
there is a tendency for increasing downward mobility for classes I, II, IIIa, and IIIb,
which is compensated by an increase in upward mobility for classes IVa and IVb. Finally,
there is a remarkable stability in the immobility rates of manual workers and an alarming
tendency for farm workers to stay in rural or urban manual working positions. It is not
possible to determine if these tendencies are provoked by economic stagnation after 1982
or if they are simple “trend-less” fluctuations. It is also difficult to determine whether the
decline of immobility in the rural sector until 1982 and the stability of immobility
afterwards results from the steady industrialization until 1982 and the economic
stagnation in the 1980’s and 1990’s. Despite fluctuations, stability and the continuous
impact of the transition into industrialism are the outstanding characteristics in mobility
chances from 1973 to 1996.
38
(3) The analyses of the impacts of the disparity between origin and destination
class distribution led to several important conclusions. First, the impact of origindestination disparity varies more among social classes than across the years studied. In
other words, the impacts of industrialization and urbanization on gross mobility is
constant from 1973 to 1996, and different for each social class.
Second, some aspects of the Brazilian patterns of gross mobility are common to
many other industrial societies. For example, lower-routine non-manual workers (IIIb),
and supervisors of manual work (V) are classes that operate as channels for upward
mobility, since they attract workers from lower destinations and at the same time send
workers to higher classes. Another common characteristic with recently industrialized
societies is that rural classes (IVc and VIIb) display high levels of self-recruitment and
send many offspring to all other classes (i.e., Costa-Ribeiro and Scalon, 2001; Hout,
1989; Breen and Whelan, 1996; Simkus, Jackson, and Treiman, 1990). Indeed, inflow
rates indicate that in each one of the class destination positions at least 20 percent of their
members have origins in the agricultural working class (VIIb). The only exception from
this rule is class I – higher-grade professionals and administrators – that has less than 10
percent of its members with origins in the rural working class.
Finally, the analyses of “structural mobility” rates indicate that the high
proportion of men with rural classes origins found in all other classes resulted from
disparities between inflow and outflow patterns, instead of from any increasing openness
in the class structure. In other words, the analyses of structural mobility patterns from
1973 to 1996 confirmed the fact that rapid industrialization and urbanization are the main
causes of the high levels of observed mobility in Brazil throughout the period studied.
39
When combined to the analyses in the two last chapters, the findings of the
present chapter suggest a striking conclusion about the Brazilian class structure: although
there are high levels of gross mobility, the major divisions within the class structure do
not change neither from 1973 to 1996, nor between generations. Despite the fact that this
conclusion is somewhat exaggerated, since from origins to destination there was a
significant increase in white-collar and blue-collar urban work, it is valid because the
major divisions in the class structure do not changed.
My descriptions lead me to agree with one of Sorokin’s (1927/1959: 153)
conjectures about social mobility patterns and stratification systems in his classical book.
In his “Social and Cultural Mobility” (1927), Soronkin suggests that while certain
barriers to mobility have been removed, one should recognize that other barriers have
become more difficult to cross. In Brazil, although the transition from a rural into an
urban society was fast and impelled high volumes of mobility, the high amounts of
workers in the most underprivileged positions changed only slightly. In other words, the
proportion of workers in unskilled working classes continues to be extremely high in
Brazil even when one compares origin and destination distributions. Of course,
industrialization in Brazil provoked a decrease in rural work, but it did not significantly
diminish manual work and lower grade non-manual work. Moreover, the fact that the
highest white-collar class positions include a large number of people in absolute terms,
although small as a proportion of the total population, suggests that economic
development increased the proportion of persons in upper white-collar class positions,
without changing major divisions within the class structure. The divisions between nonmanual, manual, and rural work, between skilled and unskilled work, and between
40
proprietary and non-proprietary classes seem to be almost unchanged in spite of fast
industrialization and urbanization that characterized the recent history of Brazil. Marx’s
([1852] 1978) idea of the class structure in “constant flux” without changing the main
class distinctions continues to be inspiring to understand the interconnection between
classes and mobility flows.
In the next chapter, I describe the degree of inequality of social mobility chances
among people coming from different class origins. These descriptions will complement
the foregoing analyses, since they determine the degree of openness or social fluidity in
Brazil from 1973 to 1996.