Do Economic and Political Changes Matter? Intergenarational Occupational Mobility in Spain

Introduction and Background

Social mobility refers to the extent of changes in an individual’s social and economic status either within the life course (intragenerational) or across generations (intergenerational). ¹ It is therefore a key factor in determining the characteristics and the performance of society, especially from an economic point of view. The importance of social mobility is also supported by additional arguments. Increased social mobility facilitates the efficient allocation of economic resources, which is positively related to a greater capacity for change, innovation, and economic growth in general. It is also linked to a more open society, with a wider range of economic opportunities for people, and greater freedom of choice.

An analysis of the intergenerational transmission processes between parents and children is one of the commonly used approaches to study social mobility, especially from an economic and sociological point of view. An intergenerational transmission analysis compares economic results between different generations. In other words, it aims to determine the extent to which the offspring’s economic outcomes are determined by the situation and circumstances of their parents. ² Empirically, what we usually observe is a high degree of correlation between the economic results of parents and their children from different perspectives (income, earning, education, and labor occupations).

Intergenerational transmission has been the subject of considerable research. ³ Intergenerational transmission processes are a consequence of the investments made by parents in their children’s education, for example, and of the resources that are transmitted: income and wealth, such as through inheritances, and social capital such as social networks and noncognitive abilities linked to personality or intellectual capacity. The intensity of parents’ transmission to their children is conditioned by the effect that some household factors exert (the parent’s education and their level of income, inheritances and household traits, among others), the social context, and other personal factors including genetic inheritance. ⁴ As the correlation is higher between parents and children, the more intense the transmission, the lower the intergenerational mobility. As further economic outcomes observed in children are explained by the characteristics and position of their parents, these results will be predetermined with greater intensity, so that social mobility will be low. ⁵ In some cases, there is a high level of overlap between the jobs that parents and children do. For example, Corak and Piraino establish a 40 percent coincidence. ⁶ Beller and Hout report a generalized range between 0.3 and 0.4. ⁷

The literature on transmission between generations considers various analytical perspectives. The most common focus on analyzing the correlation between parents and children, mainly in terms of income, earning, level of education, and labor occupation. ⁸ However, the analysis of intergenerational transmission of labor occupation has several advantages compared to other alternatives. First, compared to income, labor occupations appear to be a better approach to the concepts of status and social privilege. ⁹ Second, labor occupations to some extent include the other perspectives, in that they usually entail some degree of remuneration and require specific educational achievements. ¹⁰ Furthermore, data about labor occupations are less subject than the alternatives to problems repeatedly pointed out in the literature. ¹¹ The first of these is the possibility of being affected by transitory shocks—meaning that information reported about income at a particular point in time does not faithfully represent the parent’s permanent level of income. The second is its variation in terms of the parents’ and children’s age. The third is the bias that income data can introduce in the analysis from a life cycle perspective. However, labor occupations remain relatively steady throughout the professional life, and as such they are a better indicator and provide better information about an individual’s long-term labor situation. Finally, taking into account how retrospective information about parents’ situation is obtained, labor occupation is much more easily remembered by the children than their parent’s income level at a particular point in time in the past.

Nowadays, there is no scholarly consensus about the source of cross-national differences in mobility. For example, Bordieu et al. analyze occupational mobility in France and the United States during the nineteenth and twentieth centuries. ¹² They conclude that the differences observed in the relative evolution of social mobility in the two countries could be explained by the different degrees of urbanization, the level of industrialization, access to education, social invigoration caused by migration, and, in particular, the effects of political and social changes on institutional barriers to the mobility. Likewise, Ferrie and Long and Ferrie study the evolution of occupational mobility in the United States, first, and in the United States and the United Kingdom, second, from 1850 in both cases. ¹³ Their results highlight factors related to changes in geographical mobility and, to a lesser extent, in access to education and some institutional modifications in the labor market as the most important factors. Yaish and Andersen use data from the International Social Survey Programme to test the main explanatory hypothesis related to contextual factors in twenty modern economies. ¹⁴ This work mentions the importance of contextual factors, especially the industrialization hypothesis, in explaining intergenerational transmission. By contrast, the relevance of economic inequality is only very moderate. Di Pietro and Urwin establish the influence of the economic or social position of parents over the labor outcomes of their children. ¹⁵ Finally, Piketty proves the relationship between social mobility and redistributive policies. ¹⁶ Novel methodological advances have improved in the estimations of precise correlations ¹⁷ and elasticities in the measurement of mobility; these mainly analyze mechanisms of causality. ¹⁸

Part of the literature on intergenerational transmissions has focused its efforts on illustrating the different gender issues. ¹⁹ In one of the seminal works in this area, Havens establishes that men and women have different degrees of occupational mobility, which is sometimes due to their unequal participation in the occupational structure. ²⁰ Stevens shows in the case of Canada that the offsprings’ occupations depend on the linkages established between gender (father–son, father–daughter, mother–son and mother–daughter). ²¹ The findings establish that occupation segregation by gender is perpetuated due to different transmissions by gender.

The classic models of occupational gender-typing leads have been updated with recent research. De Graaf and Kalmijn estimate occupational transmission in the Netherlands while splitting men and women. ²² The transmission from fathers is usually stronger for sons than for daughters, and transmission from mothers is usually weaker than from fathers. Korupp et al., ²³ using data for the German economy, show that father–son transmission is usually stronger than other types of family transmissions, but although mother–daughter occupational transmission tends to be weaker, it is also significant, which reinforces the ideas of Stevens and Boyd. ²⁴ As established by the authors, this contradicts the model of status attainment of Blau and Duncan based mainly on the combination of the parents’ status. ²⁵ Their findings show that same-gender transmissions are stronger than cross-gender transmissions, and the most significant links are always to the fathers rather than to the mothers. There are even new studies that explore intergenerational transmission not only from parents but also from parents-in-law. ²⁶

In the case of Spain, De Pablos and Gil confirm the existence of a gender bias by focusing their analysis on educational and occupational mobility. ²⁷ Some explanatory factors for these different patterns of transmission by gender could be the higher wages that fathers generally earn, the different degree of socialization of sons and daughters, the different support for education by gender, and/or the persistence of gender roles, among others.

As well, the seminal work of Carabaña had analyzed the intergenerational mobility using the matrices of transition and log-linear models for Spain. ²⁸ From this initial study, an important subsequent analysis has been developed by Salido, Perales and Herrera-Usagre, Martínez-Celorrio and Marín-Saldo, Fachelli and López-Roldán, and De Pablos and Gil. ²⁹ To a greater or lesser extent, this research shows a medium social mobility in our country in comparison with other European countries, the importance of including and considering women in this type of analysis, and a slight increase of social mobility in the 1970s.

Hypothesis and Motivation

In relation to occupational intergenerational transmission, the recent literature highlights the possible explanatory role played by contextual factors in terms of the processes of intergenerational transmission. ³⁰ The main idea is that economic, political, and institutional conditions could influence social mobility and intergenerational transmission. Yaish and Andersen summarize up the current main ideas supporting this hypothesis ³¹ :

First, the industrialization and tertiarization thesis. Economic development, embedded in the concept of structural change, stimulates occupational mobility between generations based on two main arguments. First, industrialization and the increase in services favor a structural change in the economy that creates new requirements in labor occupations. Second, industrialization favors increased urbanization, more importance for education, and a more intense geographical mobility for individuals. The study by Motiram and Singh shows that there is a high degree of intergenerational persistence among agricultural workers. ³² In this context, the extensive tertiarization of the Spanish economy over the last decade should encourage an increase in intergenerational occupational mobility. Tertiarization is also accompanied by a professionalization of labor, involving the existence of more qualified labor occupations. Occupational transmission should be higher in the later historical stages in our results.

In the context of previous ideas, the aim of this article is twofold. First, we aim to study the level of intergenerational mobility and its evolution in Spain from the 1960s to the present. Second, our objective is to connect the far-reaching political, social, and economic transformation that took place in Spain during this period with the evolution of social mobility. In order to do so, the processes of labor occupations transmission between parents and children will be analyzed, emphasizing the importance of belonging to various historical stages. We will use microdata on poverty transmission from the specific module of the Spanish Living Condition Survey (2011).

This article makes three main contributions. First, we study the case of Spain, in which the recent political and socioeconomic transformations probably have more historical relevance in comparison with the European context. Second, our research extends the Spanish case into the international literature, which has so far focused exclusively on an analysis of cases from the United States, the United Kingdom, and France, thereby enabling us to check whether previous international results are also observed in Spain. Third, unlike the limited previous studies concerning the Spanish case, ³⁵ our research is carried out based on a microeconometric analysis, so that a Heckman ordered probit model and a Heckman probit model are developed to check the previous hypothesis and obtain outcomes.

In relation to these ideas, our work emphasizes that intergenerational transmission does not only depend on the evolution of history, in the sense of the simple passage of time, because some historical stages may be more crucial than others due to the creation of a climate favorable to social mobility. We attempt to adapt the ideas of Long and Ferrie about the importance of distinguishing between different historical periods in the analysis of labor mobility. ³⁶ Our analysis also shows the relevance of gender bias. Accordingly, the estimations will be recalculated by gender, endeavoring to show that men and women present different patterns in terms of intergenerational transmission and that in Spain, fathers’ (mothers’) exert an especial influence on sons (daughters). Our article proceeds as follows. The second section reviews the main characteristics of the political, social, and economic evolution in Spain over the last half-century. The third section describes the methodological approaches and data used. The outcomes are analyzed in the fourth section. The article concludes by setting out and commenting on the main conclusions (the fifth section).

Political, Economic, and Social Transformations in Spain since the Mid-twentieth Century

Spain has experienced huge transformations over the last half-century. They are probably unparalleled in other European countries. Spanish society has changed radically from both a political and an economic point of view. ³⁷ As mentioned above, our aim is to ascertain the characteristics linked to the great transformation that took place in Spanish society during the critical period from the 1960s to the beginning of the twenty-first century. Accordingly, the analysis period will be divided into three sections, based on central historical milestones: Franquismo (1960–1975) related to the Franco regime, the political and economic Transition (1976–1986), and the integration in Europe (1987–2000). In order to carry out the analysis and according to the literature mentioned above, we will consider six different analytical perspectives: changes in the political system, economic transformations, the evolution of the institutional performance of the labor market, structural changes, and the transformation of the tax system and educational system. Rather than providing a detailed description, our aim is to stress the main highlights in Spain’s recent history from the aforementioned perspectives, while attempting to limit our focus on the factors that influence generational transmission processes to the greatest extent. ³⁸

From a political perspective, there is no doubt that the main feature in Spain’s recent history has been the country’s evolution from an authoritarian regime to a democratic one. Francisco Franco imposed a dictatorship—“Franquismo”—between his victory in the Spanish Civil War in 1939 and his death in 1975. Democracy was achieved through a political process named “Transition” which, according to most historians, ended in 1986, when the Accession Treaty of Spain to the European Economic Community was signed—“Europe.”

In terms of transformation of the economic system, the Franco dictatorship imposed an autarchic economic system, isolated from the outside world, highly regulated, and based on a strategy of national production exclusively oriented “inwards.” ³⁹ The economic stagnation that characterized this economic system led to its partial temporary reform through the National Economic Stabilization Plan in 1959, which led to some economic growth in Spain, which was boosted by the liberalization of imports and increased economic opening to other countries. However, the threat that these changes implied to the political regime restricted reforms to the application of a succession of Development Plans after 1964, which meant that the system returned to its original philosophy. In simple terms, Franquismo involved an economic system that was autarchic, ⁴⁰ intensively regulated, without a modern tax system, with a very rigid and regulated labor market and high levels of income inequality. ⁴¹ This economic system, defined as corporative capitalism, suffered from a very rigid performance with high levels of intervention and limited capacity to adjust to the successive oil shocks of 1973 and 1979. With the dictator’s death, these factors led to the disintegration of Franquismo and its replacement with a new democratic system, but this transition was accompanied by a severe economic crisis. Spain was only able to apply a correct Economic Policy to address the consequences of the economic crisis after the signing of the national agreement called the “Pactos de la Moncloa” in 1977. Furthermore, the “Pactos de la Moncloa” made the reform of the previous Spanish economic system possible: deep-seated reforms to the tax and financial system, labor relationships, and the Public Administration. These transformations can be summarized by saying that they involved the substitution of a very regulated and intervened economic system with another clearly market-oriented one. These processes made possible a new economic beginning, which ended with Spain’s entry into the European Economic Community in 1986. The consequence was economic growth and socioeconomic transformation that was the cornerstone for a long process of convergence with and assimilation of European standards. The initial rigidity, the economic crisis of the 1970s, and the change of regime in politics made a far-reaching modification of labor market regulations possible. A new Industrial Relations Law was approved in 1976, introducing the concept of unfair dismissal into the labor regulations. The trade unions were recognized in 1977, as well as the right to strike and collective bargaining. The national labor statute, the “Estatuto de los Trabajadores,” was enacted in 1980, defining a new framework for labor relationships based on permanent labor contracts, the limitation of fixed-term employment, and processes of dismissal linked to high costs of firing. The result was that the “Transition” made the shift to a modern institutional framework for labor relationships possible, but it was not able to overcome the traditional rigidity of the Spanish labor market. ⁴² A deep reform of the institutional framework of the Spanish labor market took place in 1984. Fixed-term labor contracts became widespread with the abolition of previous principle of causality. The objective was to bring flexibility into the labor market, as a way to address the very high unemployment rates in the Spanish economy (up to 20 percent). The labor reform of 1984 undoubtedly achieved its objectives to a large extent, taking into account that the temporary rate increased dramatically, reaching levels of 30 percent of the active population.

However, the new situation was not without problems. A high temporary rate became chronic and the source of declines in productivity. ⁴³ As a result, after 1994, new labor reforms introduced institutional regulation in an attempt to extend flexibility while at the same time trying to reduce levels of temporary employment. Successive reforms that were carried out in Spain (1994, 1997, 2001, 2006, 2010, and 2012) led to some reduction in the legal opportunities for the use of fixed-term labor contracts, and a moderate reduction in the cost of dismissals linked to permanent employment. After reviewing these historical changes in the Spanish economy, it is clear that the greatest institutional transformations took place after the end of the Franco regime. According to the hypotheses established in the introduction, this historical stage should therefore be one of the most important periods in terms of social mobility.

Structural change (Figure 1) has been very far-reaching since “Franquismo” in Spain, as can be seen above all in the dramatically reduced importance of agriculture, in which total employment fell by half, and the associated increase in service activities. In 1993, the active population in Spain working in those intangible economic activities exceeded 60 percent of total employment.

OPEN IN VIEWER

Figure 1.

Changes in sectoral distribution of employment in Spain. Percentages.^a Methodological break of series in 1996. Source: Survey of active population Encuesta de Población Activa (EPA) Spanish statistical institute Instituto Nacional de Estadística (INE).

Finally, in Spain, the educational system has been deeply transformed. An intense process of alphabetization was developed during “Franquismo,” but the main change was afterward with the increase of high education between population. This trend was really important between women, actually, approximately the 30 percent of women have higher education. Caparrós shows that people with higher education have a higher probability of mobility up and are the leaders of social mobility in Spain. ⁴⁴ Gil-Hernández et al. establish similar issues emphasizing the importance of education in the cohorts born in the 1970s. ⁴⁵

From the arguments above, it is possible to highlight the main characteristics (political, economic, and structural labor market) of the periods mentioned:

First, the historical stage of Franquismo (1960–1975) could be defined by an autarchic and state-regulated economic system, a rigid and intervened labor market, with agriculture considerably important. The starting point of our analysis is therefore characterized by rigidness and few opportunities for change. It was a society with initial but limited social mobility.

According to the literature, we would expect these elements to have contributed to an increase in intergenerational transmission, based on various arguments:

Unlike authoritarian regimes, democratic regimes should promote social mobility. Furthermore, opening up from an economic system subject to intervention could increase opportunities for mobility.

OPEN IN VIEWER

Figure 2.

Historical changes in labor occupations distribution. Percentages. Methodological break of series in 1994 (change in national classification of occupations) and 1996 (methodological break). WCHS = white-collar high skill, WCLS = white-collar low skill, BCHS = blue-collar high skill, and BCLS = blue-collar low skill. Source: Survey of active population EPA Spanish statistical institute INE.

In short, our hypotheses are explained by the following arguments:

Methodology and Data

This analysis uses data from the Spanish Living Conditions Survey of 2011 (SLCS-2011). This database includes a specific module designed to analyze poverty transmission processes. The module contains a specific set of questions about some aspects related to parents and thereby makes information available about the interviewees (children) and their fathers and mothers. ⁴⁶ This information concerns their level of education, employment, and the parents’ labor occupation when the children were fourteen years old. This special module therefore enables us to study the transmission processes between parents and children in terms of labor occupation. Unfortunately, this module was not continued in subsequent years. In fact, it is the only microdata available on this subject in Spain. However, some shortcomings associated with this database must be noted.

With regard to labor occupations, the SLCS-2011 tabulates the information according to the National Occupations Classification, with a disaggregation of two digits for the interviewees (sons and daughters) and with a one-digit disaggregation for the parents. In the former case, the labor occupation reported by the interviewees is associated with their current job or the occupation performed in the most recent job, depending on their employment situation (employed or unemployed). For the parents, it is the father/mother’s labor occupation as remembered by the interviewee when he or she was fourteen years old.

This information is used to establish a variable with ten categories: nine for common labor occupations ⁴⁷ and a final category that for the parents includes individuals with no employment (unemployed) or out of the labor market (inactive). ⁴⁸ In our opinion, this latter category may be important for including frequent inactivity shown by some specific groups, particularly mothers, during certain historical periods, mainly during Franquismo. ⁴⁹ The progressive increase in the female labor supply is one of the distinguishing features of the Spanish labor market and shows the important transformation it has undergone in its recent history. The analysis does not consider people working in the Army, due their high specificity and the difficulty of establishing its order within an occupational hierarchy.

Our final interest is to determine the influence exerted by the political, social, and institutional transformations that occurred in Spain on intergenerational transmission. To do so, we divided our database, SLCS-2011, into three historical periods: Franquismo (1960–1975), the Transition (1976–1986), and Europe (1987–2000) based on the time when information about the parents is reported. The objective is to examine and compare the intergenerational transmission processes that took place in each of these periods. In specific terms, the strategy was as follows. The SLCS-2011 provides information about individuals aged between twenty-five and fifty-nine who report information about their parents when they were fourteen years old. If we subtract fourteen years from the interviewee’s age in 2011, we will know for how long the parent–child transmission happened for each interviewee. We then subtract 2011, the year of the survey, from the previous number and thereby obtain the year that applies to the parents’ situation reported. Finally, we classify interviewees according to the time when the transmission is reported, from 2000 to 1966, sorted as follows:

People aged between fifty and fifty-nine in 2011. Their SLCS-2011 reported transmission took place during the Franquismo period (1960–1975).

The sample includes a total of 16,630 cases with an unweighted period; 4,858 referred to Franquismo; 5,927 to the Transition; and 5,845 to Europe. It should be remembered that this is not the usual way to study the evolution of generational transmission, which is normally based on the children–parent’s data for each year considered. However, in our opinion, the unavailability of this kind of data justifies the use of the strategy described above. Our assumption is that splitting by the age of individuals who reported their parent’s occupational circumstances when they were fourteen in 2011 enables us to examine the economic transmission they received from parents who performed this transmission in different historical contexts. The age of the children remembering their situation when they were fourteen years old leads us to transmissions that occurred at different times and were influenced by very different contextual factors. However, we are aware that this approach raises several difficulties. First, the fact that the children’s occupations were all observed in 2011 suggests that they are affected by the structural transformation that happened between 1960 and 2000. Second, there is a life cycle effect on occupations, meaning that older people have a higher probability of performing a high occupation than younger people, solely due to the effect of age. Third, there may be some selection problems linked with the female low labor participation rate, especially during the Franquismo period (see Table 1) and people who are unemployed or out of the labor market.

OPEN IN VIEWER

Table 1.

Descriptive Analysis of Occupations Transmission between Sons and Parents. Thousands of People and Percentage.

	Reference Person		Father		Mother
Labor Occupations	Thousands of People	Percent	Thousands of People	Percent	Thousands of People	Percent
Elementary occupations	3,610.8	16.5	3,075.8	14.3	1,738.9	7.6
Agricultural workers	603.5	2.8	3,109.9	14.4	641.3	2.8
Service and commercial	3,998.6	18.3	2,005.7	9.3	1,465.6	6.4
Construction and manufacturing workers	2,818.5	12.9	4,454.5	20.7	513.0	2.3
Operators	1,805.2	8.2	2,594.8	12.0	232.0	1.0
Administrative	2,603.9	11.9	1,323.4	6.1	581.2	2.6
Technicians and associate professionals	2,327.6	10.6	1,809.2	8.4	260.7	1.1
Professionals	3,018.7	13.8	1,128.1	5.2	679.2	3.0
Directives	1,095.2	5.0	1,313.9	6.1	255.0	1.1
Out of employment/labor marketa	—	—	720.5	3.3	16,400.0	72.0
Total	21,882.1	100.0	21,535.7	100.0	22,766.9	100.0

Source: Spanish Living Conditions Survey of 2011.

^aPeople without information (no response) in the survey have been removed, so that it is only referred to unemployed (out the employment) or not in the labor force (out of the labor market).

Some decisions have been taken in order to address these shortcomings. First, we run estimation including agricultural employment and its related labor occupations, trying to capture the importance of the sectoral and occupational transformations that have happened in the Spanish economy. Second, we limited our target population to twenty-five to fifty-nine years. Third, and most importantly, our models have been estimated including the Heckman selection filter to address the selection bias. We therefore estimate two equations. In the first equation, we estimate the probability of being active versus being inactive. The variables that we used for this estimation are age (as a continuous variable), being female (dummy), level of education (dummy), existence of children in the household (dummy), and whether the situation of the home is bad (they have difficulties ⁵⁰ ). Including age as independent variable can reduce the possible bias mentioned above, conserving older people having a higher occupation and therefore greater mobility. With the results of this equation, a second equation is weighted and the results are corrected for not participating in the labor market and having no occupation (unemployed or out of the labor market). Table A1 in the Appendix contains a description of the independent variables. The selection bias related to labor participation (among both women and people without an occupation) is to some extent resolved. The second equation is an ordinal probit on the occupational categories of the children and shows the occupational progress weighted by the participation estimated previously. Using the information from SLCS-2011 described above, we carried out several complementary analyses. The first objective was to ascertain the determinants of the transmission processes in labor occupations from parents to children, considering the possibility of different historical periods leading to important statistical differences between generations. The second aim was to quantify the intensity of intergenerational mobility in Spain and to observe its evolution over time. Finally, the analysis emphasizes the gender perspective, as the models for sons and daughters are separated.

In order to achieve the first objective (determinants of transmission), two Heckman ordered probit models, as explained above, are estimated in order to determine the probability of an individual working in a certain labor occupation. According to Carmichael, ⁵¹ labor occupations could be differentiated and ordered in terms of their associated wages and labor status. The estimate of a Heckman ordered probit model is therefore adequate for analyzing the transmission of labor occupations between parents and children. The order of labor occupations enables us to check whether some occupational improvement between parents and children has taken place. Following the methodology used by Carmichael, ⁵² we estimate a Heckman ordered probit model in which the categories of the dependent variable (the interviewees’ labor occupations) have been ordered according as shown in Table 2. This order reflects both the wages and skills levels associated with each occupation (from lower to higher skilled). ⁵³ Based on this order, the results are interpreted directly, reflecting the probability of reaching a higher occupation in an established order; therefore, achieve occupations’ advance or achievement. When estimating this Heckman ordered probit model, the dependent variable is the labor occupation of the interviewee. The independent variables are the father’s and mother’s labor occupations ⁵⁴ and a set of other pieces of control information: personal variables (gender, nationality, and level of study), household (marital status), and variables linked to their family background (father or mother born in a foreign country and the household’s economic situation). ⁵⁵ The model is estimated first for the entire sample and then separately for men and women (son and daughter) to ascertain different patterns of mobility by gender. The regional dummies are considered controls.

OPEN IN VIEWER

Table 2.

Average Annual and Monthly Gross Wage per Worker by Labor Occupations.

Labor Occupations	Average Annual Gross Wage per Worker (Euros)a	Average Monthly Gross Wage per Worker (Euros)
Elementary occupations	10,581.8	895.7
Agricultural workers	11,697.4	1,029.5
Service and commercial	13,375.6	1,059.5
Construction and manufacturing workers	16,319.0	1,333.5
Operators	18,510.3	1,390.6
Administrative	19,309.9	1,414.9
Technicians and associate professionals	22,532.5	1,613.3
Professionals	29,232.8	2,038.5
Directives	35,347.4	2,635.9

Source: Spanish Wage Structure Annual Survey of 2011.

^aThe year of reference is 2010, the previous year of survey.

Table 1 shows some descriptive data about labor occupation transmission between parents and children. SLCS-2011 clearly shows how some significant changes have happened. While in the case of the fathers, the occupational distribution was a consequence of a historical productive structure biased toward manufacturing and agriculture, in the case of the offspring, the occupations are more related to service labor occupations. The greatest increase is for “service and commercial workers,” a category with almost double the share compared to the fathers (18.3 percent vs. 9.3 percent). For the mothers, there was a higher percentage for out of employment/labor market (72 percent), associated with the higher levels of absence of women in Spain’s recent history. Among mothers who were working, the highest concentration was observed in “elementary occupations” and “service and commercial workers.”

As regards the second objective (the intensity of the transmission), a set of additional Heckman probits models were estimated in order to analyze the probability of occupational advance between the labor occupations of the interviewees and their parents. The dependent variable is defined by whether the labor occupations of the sons/daughters were higher than their parents or otherwise (lower or equal), according to the order above. The independent variables are the same as in the models explained above, plus a set of regional dummies. Using this estimate, we will calculate the predicted probabilities for an individual in terms of occupational advance compared with his or her parents, both for the whole period and for each subperiod. These models obviously do not consider the possibility that parents do not work or are not included in the labor market because it does not allow a labor comparison between the parents’ occupation and the occupation of the children.

As a summary of transition tables, Table 3 displays descriptive information related to this comparison. On the basis of the SLCS-2011 data, the levels of occupational matching are moderate (around 20 percent of the children have the same labor occupation as their parents, although the percentage is slightly higher when the comparison is only with the mother). However, this does not fall within the consensus that the occupations of fathers and sons ranges between 0.3 and 0.4 on average. By gender, Fachelli and López-Roldán estimate a permanence of 0.23 for men and 0.17 for women. ⁵⁶ This may be indicative of the importance of analyzing the Spanish case, as an example of an intense and quick social transformation where there has been significant mobility. However, between 24.7 and 33.7 percent of the children experienced an occupational regression compared to their parents and between 45.4 and 52.8 percent present an improvement in their occupational status. Unfortunately, information about the mothers’ labor occupations is scarce because many of them were traditionally excluded from the labor market, especially among the older generations. ⁵⁷ The estimates for mothers must therefore be interpreted taking this absence into account. Based on this initial descriptive analysis, it seems that occupational mobility has to some extent been high in the Spanish economy along the period considered.

OPEN IN VIEWER

Table 3.

Occupational Matching.

Children versus Parents Occupations	Father			Mother
	Equal (%)	Lower (%)	Upper (%)	Equal (%)	Lower (%)	Upper (%)
Elementary occupations	4.5	11.3	—	6.5	7.8	—
Agricultural workers	1.5	0.8	0.4	1.2	0.5	0.6
Service and commercial	2.6	10.4	4.6	5.9	6.9	7.3
Construction and manufacturing workers	4.1	3.2	5.2	1.5	2.5	7.5
Operators	1.9	1.4	5.3	0.3	1.7	6.2
Administrative	1.6	3.2	7.5	1.4	2.5	7.1
Technicians and associate professionals	1.6	1.7	7.7	0.6	1.9	8.3
Professionals	2.3	1.6	10.3	4.4	0.9	11.3
Directives	0.8	—	4.5	0.6	—	4.5
Total	100	100	100	100	100	100
Coincidence	20.9			22.5
Occupational regression	33.7			24.7
Occupational advancement	45.4			52.8
Total	100			100

Note: Summary of transition matrixes. Percentages on the diagonal, above and below.

Source: Spanish Living Conditions Survey of 2011.

Outcomes

Table 4 shows the outcomes of different estimations of Heckman ordered probit models of the probability of the interviewees having a high labor occupation in the labor occupation hierarchy (determinants of transmission). According to the methodology described above, the children’s labor occupations are ordered from “directives” (the highest category) to “elementary occupations” (the lowest category). First, the model was estimated for the whole sample and then it was differentiated by gender (sons and daughters).

OPEN IN VIEWER

Table 4.

Heckman Ordered Probit Models: Determinants about Children’s Labor Occupation.

	Total		Sons		Daughters
	Coefficient	Margins (dy/dx) at Means (Outcome = 9)	Coefficient	Margins (dy/dx) at Means (Outcome = 9)	Coefficient	Margins (dy/dx) at Means (Outcome = 9)
Father’s occupation
Elementary occupations	−0.147***	−0.012	−0.161**	−0.018	−0.209**	−0.006
Agricultural workers	−0.081	−0.007	−0.168**	−0.019	0.026	0.001
Service and commercial	0.130**	0.011	0.058	0.007	0.221**	0.006
Construction and manufacturing workers	−0.001	0.000	−0.028	−0.003	−0.015	0.000
Administrative	0.166***	0.014	0.131	0.015	0.190*	0.005
Operators	RC		RC		RC
Technicians and associate professionals	0.121**	0.010	0.061	0.007	0.191**	0.005
Professionals	0.246***	0.021	0.179**	0.020	0.344***	0.009
Directives	0.342***	0.029	0.344***	0.039	0.320***	0.009
Out of employment/labor market	0.055	0.005	0.020	0.002	0.031	0.001
Mother’s occupation
Elementary occupations	0.000	0.000	0.125	0.014	−0.115	−0.003
Agricultural workers	0.077	0.007	0.443*	0.050	−0.300	−0.008
Service and commercial	0.104	0.009	0.164	0.019	0.064	0.002
Construction and manufacturing workers	0.093	0.008	0.169	0.019	0.063	0.002
Administrative	0.323**	0.027	0.394*	0.044	0.306	0.008
Operators	RC		RC		RC
Technicians and associate professionals	0.232	0.020	0.264	0.030	0.346	0.009
Professionals	0.466***	0.039	0.635***	0.072	0.357	0.010
Directives	0.439**	0.037	0.613*	0.069	0.279	0.008
Out of employment/labor market	0.122	0.010	0.296	0.033	−0.057	−0.002
Historical period
Europe	RC		RC		RC
Transition	0.112***	0.009	0.133***	0.015	0.098*	0.003
Franquismo	0.194***	0.016	0.236***	0.027	0.170***	0.005
Gender
Men	RC		RC		RC
Women	−0.238***	—	—	—	—	—
Marital status
Single	RC		RC		RC
Married	0.118***	−0.020	0.181***	0.020	0.019	0.001
Others	−0.017	0.010	0.111	0.012	−0.087	−0.002
Level of study
Primary	RC		RC		RC
Secondary: first stage	0.171***	−0.001	0.153**	0.017	0.393***	0.011
Secondary: second stage	0.436***	0.014	0.404***	0.046	0.735***	0.020
Vocational training	0.389***	0.037	0.417***	0.047	0.615***	0.017
University	1.276***	0.033	1.149***	0.130	1.759***	0.048
Nationality
Spanish	RC		RC		RC
Foreign-born	−0.213**	0.108	−0.162	−0.018	−0.318**	−0.009
Foreign-born father
No	RC		RC		RC
Yes	−0.299***	−0.018	−0.149	−0.017	−0.492***	−0.013
Foreign-born mother
No	RC		RC		RC
Yes	−0.201*	−0.025	−0.214	−0.024	−0.223	−0.006
Selection equation: participation in labor market
Age	0.006***		0.014***		0.001
Women	−0.491***		—		—
Secondary: first stage	0.335***		0.303***		0.383***
Secondary: second stage	0.593***		0.517***		0.688***
Vocational training	0.643***		0.505**		0.824***
University	0.967***		0.744***		1.156***
Children at home (yes)	0.617***		0.710*		0.682**
Household economic situation (bad)	−0.238***		−0.301***		−0.168***
Constant	−0.727***		−1.013***		−1.154***
/cut1	−0.774***		−0.633***		−0.363
/cut2	−0.638***		−0.415*		−0.304
/cut3	0.055		0.111		0.678***
/cut4	0.394**		0.659***		0.743***
/cut5	0.677***		1.089***		0.824***
/cut6	1.110***		1.358***		1.520***
/cut7	1.559***		1.877***		1.894***
/cut8	2.467***		2.568***		3.189***
/athrho	−0.283***		−0.183**		−0.127**
ρ	−0.276		−0.181		0.000
Number of observations (unweighted)	14,520		6,979		7,541
Wald χ2 (34)	1,181.3		661.0		939.2
Probability > χ2	0.000		0.000		0.000
Log pseudolikelihood	−38.400.000		−21.000.000		−15.900.000

Note: *, **, and *** indicate significance level at 90%, 95%, and 99%, respectively. Person of reference is made up with omitted categories. Robust errors. Regional dummies are included. RC = reference category.

Source: Spanish Living Conditions Survey of 2011.

It can be observed that the positive influence is higher than the negative one, for both the fathers and the mothers. In other words, the fathers and mothers’ labor occupation exerts an especially improving effect on the occupational status of their children. Rather than reducing the probability of working in a high labor occupation, the effect results in increasing the probability of engaging in this kind of labor occupation. The coefficients associated with labor occupations carried out by fathers and mothers therefore show the intergenerational effects and labor mobility. If the fathers or mothers worked in administrative positions or above, the probability of the offspring achieving an occupational improvement is positive and significant. Negative influences are restricted to the case of the father having had “elementary occupations.” The most positive and significant influence from fathers and mothers comes from “professionals” and “directives.”

Other relevant information comes from the other explanatory variables. Being a woman reduces the probability of working in a high occupation. However, the level of education has a positive relationship with this probability.

The historical period variables provide interesting information. If the son or daughter reported information dating from the Franquismo period, their probability of engaging in a skilled labor occupation is higher than in the Europe period (the reference category). The analysis points to the same effect, but to a lesser extent, for the Transition period. Both coefficients are positive for the reference category; this may indicate intense mobility in both periods, but which is slightly higher during Franquismo. The passing of time has a positive effect in terms of occupational mobility, and Franquismo is a fundamental historical period. According to the theory above, the high level of tertiarization and advances in professionalization in the occupational structure could be behind these results. In opinion of the authors, occupational classifications during the Franco period were concentrated in lower occupations; there is therefore a greater probability of them being surpassed by children. This can be tested indirectly in Table 1 in which the parents present a high concentration in occupations related to agriculture and manufacturing. In the Spanish case, and according to the international literature and De Pablos and Gil, ⁵⁸ intergenerational transmission is biased by gender, meaning that the relationship between fathers and sons is usually more intense than alternative parent–child combinations. Nevertheless, although minor, the influence of the mother is significant, as demonstrated by Rosenfeld. ⁵⁹ The difference in the Spanish case is that it is only significant for sons and not for daughters. In the case of sons, “elementary occupations” and “agricultural workers” have a negative influence, although with a reduced significance, while “professionals” and “directors” have a positive influence from both fathers and mothers. In the case of daughters, the father’s influence is extensive with more or less the same patterns, but the influence of mothers is very limited, as previously established. The influence of historical periods is positive for both genders, sons and daughters, and the Franquismo period is once again decisive.

Table 5 summarizes the results for sons, daughters, and both genders based on the associated marginal effects. Intensity of transmission in the total population is 3.9 percent. The bigger transmission among sons (5.6 percent) is apparent and is significantly higher than for daughters (1.0 percent). The gender bias is intense in the Spanish case.

OPEN IN VIEWER

Table 5.

Marginal of Heckman Ordered Probit Models about Children’s Labor Occupation.

	Margins (dy/dx) at Means (Outcome = 9)	P > z	Significance
Total	.039	.000	***
Sons	.056	.000	***
Daughters	.010	.000	***

Note: *** Significance at level of 99.

Source: Spanish Living Conditions Survey of 2011.

The previous model is estimated again but now for each historical period separately. Table 6 only shows the marginal effects obtained for each historical period as a reflection of the intensity of transmission (see Table A2 in the Appendix for the whole estimation). These results can be interpreted independently without a reference category and provide better knowledge of mobility. The magnitude of intergenerational transmission increased between Franquismo (3.9 percent) and Transition (4.5 percent), but declined to Europe (3.3 percent), albeit only slightly. Accordingly, “Transition” was a period in which social mobility in Spain declined in spite of the huge social, economic, and political changes observed. However, the effects of these contextual factors seem to increase if they are compared with what happened in “Europe” in terms of economic, political, and social conditions. Intergenerational transmission first increased (less mobility) and then decreased (more mobility). The patterns of mobility are consistent with the development of service industries, the increase in skilled occupations, and in short, the flexibilization of the labor market.

OPEN IN VIEWER

Table 6.

Marginal of Heckman Ordered Probit Models about Children’s Labor Occupation for Historical Period.

Historical Period	Margins (dy/dx) at Means (Outcome = 9)	P > z	Significance
Europe	.033	.004	***
Transition	.045	.000	***
Franquismo	.039	.000	***

Note: *** Significance at level of 99.

Source: Spanish Living Conditions Survey of 2011.

Turning now to our second objective, Table 7 shows the outcomes of estimated Heckman probit models for the probability of interviewees having a higher labor occupation than their parents or, on the contrary, working in an identical or lower labor occupation (an overpassing model). Based on these results, Table 8 shows the predicted probabilities of a child working in a higher labor occupation (occupational advance) than his or her parents, differentiating between the parents’ gender. Observing the estimated coefficients, the children are more likely to surpass the father’s occupation in all historical periods, but this pattern has decreased over time. It seems that the occupational improvement compared to the father’s situation has been losing strength. The coefficients for mother–child transmission are not significant for any of the historical periods. The absence of women from the labor market and therefore of data could perhaps be the reason for these results. The marginal effects (Table 8) point to a lower probability of advances in occupational transition, with 34.8 percent compared to the fathers and 24.5 percent compared to the mothers. However, the historical evolution of the Spanish labor market establishes that progress compared to mothers has occurred in terms of the participation rate rather than occupational advances.

OPEN IN VIEWER

Table 7.

Heckman Probit Models about Occupational Advance in Relation to Parents.

	Occupational Advance in Relation to Father		Occupational Advance in Relation to Mother
	Coefficient	Margins (dy/dx) at Means	Coefficient	Margins (dy/dx) at Means
Historical period
Europe	RC		RC
Transition	.181***	.067	0.029	.009
Franquismo	.278***	.103	−0.103	−.031
Gender
Men	RC		RC
Women	−.224***	−.083	−0.302***	−.091
Marital status
Single	RC		RC
Married	.074*	.027	0.163**	.049
Others	−.068	−.025	−0.099	−.030
Level of study
Primary	RC		RC
Secondary: first stage	.196***	.073	0.407***	.123
Secondary: second stage	.321***	.119	0.556***	.168
Vocational training	.249	.092	0.407	.123
University	.823***	.304	0.977***	.294
Nationality
Spanish	RC		RC
Foreign-born	−.122	−.045	−0.002	−.001
Foreign-born father
No	RC		RC
Yes	−.058	−.021	−0.259	−.078
Foreign-born mother
No	RC		RC
Yes	−.231*	−.085	−0.245	−.074
Selection equation: participation in labor market
Age	.006***		−0.011***
Women	−.479***		−0.349***
Secondary: first stage	.330***		0.444***
Secondary: second stage	.592***		0.768***
Vocational training	.649***		0.824***
University	.958***		1.108***
Children at home (yes)	.556**		1.132***
Household economic situation (bad)	−.255***		−0.207***
Constant	−.673**		−1.500***
/athrho	.625*		0.871*
ρ	.555		0.702
Number of observations (unweighted)	14,401		7,780
Wald χ2 (34)	404.9		227.2
Probability > χ2	0.000		0.000
Log pseudolikelihood	−21.400.000		−9.410.927

Note: ***, ** and * Significance at level of 99, 95 and 90% respectively; RC = reference category.

Source: Spanish Living Conditions Survey of 2011.

OPEN IN VIEWER

Table 8.

Marginal of Heckman Probit Models about Occupational Advance in Relation to Parents.

Occupational Advance	Margins (dy/dx) at Means	P > z	Significance
In relation to father	.348	.000	***
In relation to mother	.245	.003	***

Note: *** Significance at level of 99.

Source: Spanish Living Conditions Survey of 2011.

The comparison with other Spanish studies as Caparrós, Fachelli, and López-Roldán and Perales and Herrera-Usagre ⁶⁰ is complicated, given that different moments are used and data selection joint with the methodology used is different too. However, there are some common findings: the social mobility used to be higher for women than men, the 1970s was a crucial stage for occupational mobility, and the social fluidity has come due to a significant educational increase and process of industrialization.

Conclusions

The aim of this article has been to examine social mobility in Spain over the last fifty years, paying special attention to the influence exerted by contextual factors (political, economic, structural, and social transformations) during this period. To that end, the occupational transmission between generations has been analyzed, to determine the influence that the occupational status of parents exerts on the occupational achievement of their children. The analysis carried out, using microdata from the Spanish Living Condition Survey 2011, has, in our opinion, provided a set of interesting outcomes and relevant conclusions.

First, according to the previous literature and historical transformations that have taken place in Spain, some improvements have occurred in social mobility over the last half-century. Intergenerational transmission has evolved from 3.9 percent during Franquismo to 4.5 percent in the Transition and 3.3 percent in Europe period. The evolution from an authoritarian, autarchic, and rigid socioeconomic and political system to an alternative democratic, flexible, and modern one has also led to an increase in social mobility and a reduction in the economic ties between parents and children. The changes in mobility are consistent with the predictions of international theories; tertiarization and an increase in professional occupations have a favorable influence on mobility. Furthermore, the Transition period was extremely determinant in terms of occupational mobility. However, in our opinion, there is a dramatic difference between the importance of the changes and the minimal improvement in social mobility. These have led to a genuine historical revolution in Spain, while their implications in terms of greater social mobility are only of a moderate magnitude. Although the analyzed cohorts are different, the importance of the 1970s in Spanish social mobility is also confirmed by other studies as Caparrós, Fachelli, and López-Roldán or Perales and Herrera-Usagre. ⁶¹

Second, the influence of parents—both fathers and mothers—on children is positive (better occupations) rather than negative. If the fathers or mothers worked in a medium or highly skilled labor occupation, the probability of their children working in a high-level occupation increases. The opposite effect applies to medium- to low-skilled fathers’ occupations.

Third, profound differences based on gender are apparent, in terms of both the gender of the parents and children. Intergenerational links appear to be stronger among males than females and from fathers to mothers. ⁶² However, in our study, we have shown that the traditional low level of labor participation of women has also had consequences for occupational mobility.

In short, our research provides empirical evidence about the influence of contexts and political and economic institutions on social mobility. These results are consistent with previous studies as Carabaña, DiPetre, and Yaish and Andersen ⁶³ and validate the importance of the historical context compared to intergenerational occupational transmission. The Spanish case clearly supports the hypotheses in the literature, while showing that parents have different roles in the processes of occupational transmission between generations. These differences and their effects, albeit with perhaps unexpected moderation, emerge very markedly in the case of Spain, which has undergone historical transformations in a relatively short period.