• Non ci sono risultati.

Family matters : a sibling similarity approach to the study of intergenerational inequality in Germany

N/A
N/A
Protected

Academic year: 2021

Condividi "Family matters : a sibling similarity approach to the study of intergenerational inequality in Germany"

Copied!
207
0
0

Testo completo

(1)

Family matters

A sibling similarity approach to the study of

intergenerational inequality in Germany

Lea Katharina Kröger

Thesis submitted for assessment with a view to

obtaining the degree of Doctor of Political and Social Sciences

of the European University Institute

(2)
(3)

European University Institute

Department of Political and Social Sciences

Family matters

A sibling similarity approach to the study of intergenerational

inequality in Germany

Lea Katharina Kröger

Thesis submitted for assessment with a view to

obtaining the degree of Doctor of Political and Social Sciences

of the European University Institute

Examining Board

Prof. Fabrizio Bernardi, European University Institute (EUI Supervisor)

Prof. Juho Härkönen, European University Institute

Prof. Anette Eva Fasang, Humboldt University Berlin

Prof. Markus Jäntti, Stockholm University

© Lea Katharina Kröger, 2021

No part of this thesis may be copied, reproduced or transmitted without prior

permission of the author

(4)
(5)

Researcher declaration to accompany the submission of written work

Department of Political and Social Sciences - Doctoral Programme

I Lea Katharina Kröger certify that I am the author of the work Family matters. A sibling

similarity approach to the study of intergenerational inequality in Germany I have presented

for examination for the Ph.D. at the European University Institute. I also certify that this is solely my own original work, other than where I have clearly indicated, in this declaration and in the thesis, that it is the work of others.

I warrant that I have obtained all the permissions required for using any material from other copyrighted publications.

I certify that this work complies with the Code of Ethics in Academic Research issued by the European University Institute (IUE 332/2/10 (CA 297).

The copyright of this work rests with its author. Quotation from it is permitted, provided that full acknowledgement is made. This work may not be reproduced without my prior written consent. This authorisation does not, to the best of my knowledge, infringe the rights of any third party.

I declare that this work consists of 68923 words.

Statement of inclusion of previous work:

I confirm that chapter 3 was jointly co-authored with Mr. Fabrizio Bernardi and I contributed 80% of the work.

Statement of language correction:

This thesis has been corrected for linguistic and stylistic errors. I certify that I have checked and approved all language corrections, and that these have not affected the content of this work.

Signature and date:

(6)
(7)

I Abstract

The intergenerational transmission of inequality is a research field that has sub-strands in several disciplines with findings that have consequences for the way we see and evaluate our society. Therefore, it is crucial to continuously update how we address questions in such an important research area. In this thesis, I study the importance of the family of origin for different areas of social inequality using a sibling design. I estimate the influence of the family on labor market success, partnership union formation, and occupational gender stratification in Germany using data from the German Socio-Economic Panel. The results show that the family plays a crucial role in the generations of social inequality over the life course. It affects the labor market attainment for different social origin groups and over and above a person's education, and it influences the timing of marriage, cohabitation, and living-apart-together unions. In addition, the gender composition of the sibling group creates inequality regarding occupational attainment within families.

Thus, this thesis provides a comprehensive view of how the family of origin is relevant to several areas of social and economic life in Germany. It discusses the implications of using a comprehensive approach to the family for further research and policy.

(8)
(9)

III Acknowledgments

Writing this thesis for me has been a journey with many ups and downs. Luckily, I had many people around me that supported me through this process, for which I am very grateful. First and foremost, I would like to thank my supervisor Fabrizio Bernardi, whose constant enthusiasm and willingness to discuss my research were truly helpful and inspiring. I would also like to thank the other members of my thesis committee, Anette E. Fasang, Juho Härkönen and Markus Jäntti. Their detailed suggestions and considerate comments were very much appreciated and helped me to improve many aspects of my thesis.

I was very fortunate in that I have not only met inspiring colleagues during my time at the EUI but also made some very good friends. I want to thank Diana for her sunny disposition and a wonderful year of living in our little attic apartment. I would also like to thank Ieva for her dark sense of humor and an equally wonderful year of living in our basement. Andris, Javier and Tobias provided additional much-needed entertainment throughout the years. I am also very grateful to have met Shpend, whose continuing support in all areas of my life has been very special to me.

This also seems like a good place to thank my family. They show a remarkable interest in all aspects of my academic life while at the same time understanding that I usually do not like to talk about any of it.

Lastly, I would like to thank Hannes. Without his encouragement and support, I would not have started this thesis and I doubt I would have finished it.

(10)

IV

1 INTRODUCTION ... 1

1.1 THE CHALLENGES OF INTERGENERATIONAL INEQUALITY... 1

1.2 THE RESEARCH GAP IN THE INTERGENERATIONAL TRANSMISSION OF INEQUALITY ... 3

1.3 SIBLING SIMILARITY IN SOCIO-ECONOMIC STATUS ATTAINMENT ... 5

1.4 METHODS – A SIBLING-BASED APPROACH TO THE ESTIMATION OF INTERGENERATIONAL INEQUALITY ... 10

1.5 STRUCTURE OF THE THESIS ... 16

1.6 REFERENCES ... 20

2 THE FAMILY OF ORIGIN EFFECT ON LABOR MARKET ATTAINMENT IN GERMANY ... 25

2.1 INTRODUCTION ... 25

2.2 THEORY ... 29

2.3 DATA AND MEASURES ... 34

2.4 ANALYTICAL STRATEGY ... 37

2.5 RESULTS ... 39

2.6 DISCUSSION ... 46

2.7 REFERENCES ... 52

2.8 APPENDIX ... 57

3 SOCIAL CLASS AND SIBLING SIMILARITY IN GERMANY ... 71

3.1 INTRODUCTION ... 71

3.2 THEORY ... 73

3.3 DATA AND MEASURES ... 79

3.4 ANALYTICAL STRATEGY ... 81

3.5 RESULTS ... 84

3.6 DISCUSSION ... 87

3.7 REFERENCES ... 91

3.8 APPENDIX ... 94

4 FAMILY OF ORIGIN AND FIRST UNION FORMATION IN GERMANY... 103

4.1 INTRODUCTION ... 103

4.2 THEORY ... 107

4.3 DATA AND MEASURES ... 116

4.4 ANALYTICAL STRATEGY ... 120

4.5 RESULTS ... 123

4.6 DISCUSSION ... 132

4.7 REFERENCES ... 138

5 GENDER OF SIBLINGS AND OCCUPATIONAL SEGREGATION IN GERMANY ... 145

5.1 INTRODUCTION ... 145

5.2 THEORY ... 150

5.3 DATA AND MEASURES ... 156

5.4 ANALYTICAL STRATEGY ... 159 5.5 RESULTS ... 161 5.6 DISCUSSION ... 169 5.7 REFERENCES ... 173 5.8 APPENDIX ... 179 6 CONCLUSION ... 185

6.1 THE LASTING IMPORTANCE OF THE FAMILY FOR DIFFERENT DOMAINS OF LIFE ... 185

6.2 IMPLICATIONS FOR THEORY AND METHODOLOGY... 187

6.3 POLICY IMPLICATIONS ... 188

6.4 LIMITATIONS AND FUTURE RESEARCH ... 190

6.5 OVERCOMING INTERGENERATIONAL TRANSMISSION OF INEQUALITY ... 192

(11)

V

FIGURE 2.1: SIBLING SIMILARITY FOR THREE DIMENSIONS OF LABOR MARKET ATTAINMENT - OVERALL ____________________ 41 FIGURE 2.2: SIBLING SIMILARITY FOR THREE DIMENSIONS OF LABOR MARKET ATTAINMENT - DESO ______________________ 42 FIGURE 2.3: SIBLING SIMILARITY FOR THREE DIMENSIONS OF LABOR MARKET ATTAINMENT – DESO CONTROLLING FOR PARENTAL SES 44 FIGURE 2.4: SIBLING SIMILARITY FOR THREE DIMENSIONS OF LABOR MARKET ATTAINMENT – DESO, CONTROLLING FOR PARENTAL SES

(FATHER AND MOTHER) ____________________________________________________________________ 64 FIGURE 2.5: SIBLING SIMILARITY FOR THREE DIMENSIONS OF LABOR MARKET ATTAINMENT – OVERALL (MEN ) _______________67 FIGURE 2.6: SIBLING SIMILARITY FOR THREE DIMENSIONS OF LABOR MARKET ATTAINMENT – DESO (MEN ) _________________67 FIGURE 2.7: SIBLING SIMILARITY FOR THREE DIMENSIONS OF LABOR MARKET ATTAINMENT – DESO CONTROLLING FOR PARENTAL SES

(MEN )_______________________________________________________________________________68 FIGURE 2.8: SIBLING SIMILARITY FOR THREE DIMENSIONS OF LABOR MARKET ATTAINMENT – OVERALL (WOMEN ) ____________68 FIGURE 2.9: SIBLING SIMILARITY FOR THREE DIMENSIONS OF LABOR MARKET ATTAINMENT – DESO (WOMEN )_______________69 FIGURE 2.10: SIBLING SIMILARITY FOR THREE DIMENSIONS OF LABOR MARKET ATTAINMENT – DESO CONTROLLING FOR PARENTAL SES

(WOMEN ) ____________________________________________________________________________69 FIGURE 3.1: SIBLING SIMILARITY BY PARENTAL EDUCATION _________________________________________________ 84 FIGURE 3.2: SIBLING SIMILARITY BY PARENTAL SOCIAL CLASS (EGP) ____________________________________________ 86 FIGURE 3.3: SIBLING SIMILARITY BY PARENTAL EDUCATION (MEN) ___________________________________________ 100 FIGURE 3.4: SIBLING SIMILARITY BY PARENTAL SOCIAL CLASS (EGP) (MEN) ______________________________________ 100 FIGURE 3.5: SIBLING SIMILARITY BY PARENTAL EDUCATION (WOMEN) _________________________________________ 101 FIGURE 3.6: SIBLING SIMILARITY BY PARENTAL SOCIAL CLASS (EGP) (WOMEN)____________________________________ 101 FIGURE 4.1: SIBLING SIMILARITY FOR THE TIMING OF FIRST UNION FORMATIONS __________________________________ 125 FIGURE 5.1: THE EFFECT OF HAVING A SIBLING OF THE OPPOSITE GENDER – CHOICE OF GENDER-ATYPICAL OCCUPATION _______ 162 FIGURE 5.2: THE EFFECT OF HAVING A YOUNGER OR OLDER SIBLING OF THE OPPOSITE GENDER – CHOICE OF GENDER-ATYPICAL

OCCUPATION __________________________________________________________________________ 162 FIGURE 5.3: THE EFFECT OF HAVING A SIBLING OF THE OPPOSITE GENDER - OCCUPATIONAL STATUS (ISEI) _________________ 164 FIGURE 5.4: THE EFFECT OF HAVING A YOUNGER OR OLDER SIBLING OF THE OPPOSITE GENDER - OCCUPATIONAL STATUS (ISEI) ___ 164 FIGURE 5.5: THE EFFECT OF HAVING A SIBLING OF THE OPPOSITE GENDER - OCCUPATIONAL PRESTIGE (SIOPS) _____________ 165 FIGURE 5.6: THE EFFECT OF HAVING A YOUNGER OR OLDER SIBLING OF THE OPPOSITE GENDER- OCCUPATIONAL PRESTIGE (SIOPS) _ 165 FIGURE 5.7: THE EFFECT OF HAVING A SIBLING OF THE OPPOSITE GENDER - OCCUPATIONAL PRESTIGE (MPS) _______________ 167 FIGURE 5.8: THE EFFECT OF HAVING A YOUNGER OR OLDER SIBLING OF THE OPPOSITE GENDER - OCCUPATIONAL PRESTIGE (MPS) _ 167 FIGURE 5.9: THE EFFECT OF HAVING A SIBLING OF THE OPPOSITE GENDER - EDUCATIONAL ATTAINMENT __________________ 168 FIGURE 5.10: THE EFFECT OF HAVING A YOUNGER OR OLDER SIBLING OF THE OPPOSITE GENDER - EDUCATIONAL ATTAINMENT ___ 168 FIGURE 5.11: THE EFFECT OF HAVING A SIBLING OF THE OPPOSITE GENDER (OVER EARLY CAREER) – CHOICE OF GENDER-ATYPICAL

OCCUPATION __________________________________________________________________________ 180 FIGURE 5.12: THE EFFECT OF HAVING A YOUNGER OR OLDER SIBLING OF THE OPPOSITE GENDER (OVER EARLY CAREER) - CHOICE OF

GENDER-ATYPICAL OCCUPATION _____________________________________________________________ 180 FIGURE 5.13: THE EFFECT OF HAVING A SIBLING OF THE OPPOSITE GENDER (OVER EARLY CAREER) - OCCUPATIONAL STATUS (ISEI) _ 181 FIGURE 5.14: THE EFFECT OF HAVING A YOUNGER OR OLDER SIBLING OF THE OPPOSITE GENDER (OVER EARLY CAREER) - OCCUPATIONAL

STATUS (ISEI) __________________________________________________________________________ 181 FIGURE 5.15: THE EFFECT OF HAVING A SIBLING OF THE OPPOSITE GENDER (OVER EARLY CAREER) - OCCUPATIONAL PRESTIGE (SIOPS)

___________________________________________________________________________________ 182 FIGURE 5.16: THE EFFECT OF HAVING A YOUNGER OR OLDER SIBLING OF THE OPPOSITE GENDER (OVER EARLY CAREER) - OCCUPATIONAL

PRESTIGE (SIOPS)________________________________________________________________________ 182 FIGURE 5.17: THE EFFECT OF HAVING A SIBLING OF THE OPPOSITE GENDER (OVER EARLY CAREER) - OCCUPATIONAL PRESTIGE (MPS)

___________________________________________________________________________________ 183 FIGURE 5.18: THE EFFECT OF HAVING A YOUNGER OR OLDER SIBLING OF THE OPPOSITE GENDER (OVER EARLY CAREER) - OCCUPATIONAL

PRESTIGE (MPS) ________________________________________________________________________ 183 FIGURE 5.19: THE EFFECT OF HAVING A SIBLING OF THE OPPOSITE GENDER (OVER EARLY CAREER) - EDUCATIONAL ATTAINMENT __ 184 FIGURE 5.20: THE EFFECT OF HAVING A YOUNGER OR OLDER SIBLING OF THE OPPOSITE GENDER (OVER EARLY CAREER) - EDUCATIONAL

(12)

VI

TABLE 2.1 SUMMARY STATISTICS ... 57

TABLE 2.2: ICC AND WITHIN AND BETWEEN VARIATION: BASELINE MODEL, DESO MODEL, PARENTAL SES MODEL ... 57

TABLE 2.3 SIBLING SIMILARITY IN INCOME (BASELINE MODEL) ... 58

TABLE 2.4 SIBLING SIMILARITY IN JOB HIERARCHY POSITION (BASELINE MODEL) ... 58

TABLE 2.5 SIBLING SIMILARITY IN OCCUPATIONAL STATUS (BASELINE MODEL) ... 59

TABLE 2.6 SIBLING SIMILARITY IN INCOME (DESO) ... 59

TABLE 2.7 SIBLING SIMILARITY IN JOB HIERARCHY POSITION (DESO) ... 60

TABLE 2.8 SIBLING SIMILARITY IN OCCUPATIONAL STATUS (DESO) ... 60

TABLE 2.9 SIBLING SIMILARITY IN INCOME (DESO, CONTROLLING FOR PARENTAL SES) ... 61

TABLE 2.10 SIBLING SIMILARITY IN JOB HIERARCHY POSITION (DESO, CONTROLLING FOR PARENTAL SES) ... 62

TABLE 2.11 SIBLING SIMILARITY IN OCCUPATIONAL STATUS (DESO, CONTROLLING FOR PARENTAL SES) ... 63

TABLE 2.12: ICC AND WITHIN AND BETWEEN VARIATION: BASELINE MODEL, DESO MODEL, PARENTAL SES MODEL (MEN) ... 66

TABLE 2.13: ICC AND WITHIN AND BETWEEN VARIATION: BASELINE MODEL, DESO MODEL, PARENTAL SES MODEL (WOMEN) ... 66

TABLE 3.1 SUMMARY STATISTICS ... 94

TABLE 3.2 SIBLING SIMILARITY IN INCOME BY PARENTAL EDUCATION ... 94

TABLE 3.3 SIBLING SIMILARITY IN JOB HIERARCHY POSITION BY PARENTAL EDUCATION ... 95

TABLE 3.4 SIBLING SIMILARITY IN OCCUPATIONAL STATUS BY PARENTAL EDUCATION ... 95

TABLE 3.5 SIBLING SIMILARITY IN INCOME BY PARENTAL SOCIAL CLASS ... 96

TABLE 3.6 SIBLING SIMILARITY IN JOB HIERARCHY POSITION BY PARENTAL SOCIAL CLASS ... 96

TABLE 3.7 SIBLING SIMILARITY IN OCCUPATIONAL STATUS BY PARENTAL SOCIAL CLASS ... 97

TABLE 3.8: ICC AND WITHIN AND BETWEEN VARIATION FOR WOMEN AND MEN BY PARENTAL EDUCATION ... 99

TABLE 3.9: ICC AND WITHIN AND BETWEEN VARIATION FOR WOMEN AND MEN BY PARENTAL SOCIAL CLASS ... 99

TABLE 4.1: THE DEGREE OF INSTITUTIONALIZATION AND EXIT/ENTRY COSTS BY UNION TYPE ... 111

TABLE 4.2: DESCRIPTIVE STATISTICS FOR THE ENTRY INTO FIRST UNION FORMATION, BY UNION TYPE (FOR THOSE THAT ENTER THE UNION) ... 124

TABLE 4.3: ASSOCIATION OF INDIVIDUAL AND FAMILY CHARACTERISTICS AND TIMING OF FIRST LAT UNION (HAZARD RATIOS) ... 127

TABLE 4.4: ASSOCIATION OF INDIVIDUAL AND FAMILY CHARACTERISTICS AND TIMING OF FIRST COHABITATION (HAZARD RATIOS) ... 129

TABLE 4.5: ASSOCIATION OF INDIVIDUAL AND FAMILY CHARACTERISTICS AND TIMING OF FIRST MARRIAGE (HAZARD RATIOS) ... 131

(13)

1

1 Introduction

1.1 The challenges of intergenerational inequality

One of the pillars of the legitimizing narrative of liberal societies is the idea that everyone should have the same opportunities in life (e.g. Boudon 1974). This notion of equality of opportunities is often interpreted to mean that those who show the same degree of effort should have the same chances to reach social positions in life. The allocation of social positions should thus be independent of the circumstances people come from (Roemer 2008; Roemer and Trannoy 2015). Equality of opportunity has been the subject of a large body of conceptual and empirical research (e.g. Breen 2004; Breen and Jonsson 2005; Ermisch, Jäntti, and Smeeding 2012; Torche 2015). Under this approach to the allocation of social positions, having some degree of inequality in outcomes is compatible with notions of fairness as, ultimately, the inequality could be attributed to individual decision-making or luck. Thus, this idea stands in contrast to the notion of equality of outcomes, which posits that differences in the rewards associated with social positions should be minimized, even if they are due to personal effort or talent (Phillips 2004).

Another important facet of the liberal paradigm is that of meritocracy, which is related to the ideal of equality of opportunity. A meritocratic order of society would imply that positions are allocated based on individual merit (Young 1958). While there are very many notions of how merit is defined, it is usually assumed to be determined by both effort and talent (Jackson 2007). The leading normative supposition of this strand of the literature is that if social origin determines social positions (meaning there is a substantial

intergenerational transmission of inequality) through channels other than individual effort and talent, meritocracy is not (yet) achieved and equality of opportunity not (yet) realized in this context. As seen from this perspective, the sociological study of intergenerational transmission of inequalities can be motivated by a need to assess to what degree the liberal notion of equality of opportunity and the meritocratic order of society are empirically realized in different contexts (such as different cohorts and different countries).

One key unit of analysis in this discussion is the family. Families are arguably one of the most important institutions of Western liberal societies. We can even say that a large part of the

(14)

2 societal order is based on (nuclear) family structures. For example, in Germany, the family enjoys special protection under the Grundgesetz.1 In line with liberal ideas of the autonomy

of the individual, much discretion is left to parents in a nuclear family to nurture, teach and raise their children as they see fit, without the direct intervention from the state, religious groups or broader parts of society. While this is, of course, an ideal view on how families function, ignoring many important societal influences on family life, it is undeniable that liberal societies afford parents the unequivocal right to further the welfare and success of their children in life.

Interestingly, while both the ideas of meritocracy and equality of opportunity and the autonomy of the family are core aspects of liberal societies, they are by no means in harmony with each other. It is one of the overarching claims of this thesis that liberal societies have so far not been able to reconcile the two categories and that societies – in particular, German society as studied in this thesis – have given precedence to the autonomy of the family over meritocracy and equality of opportunity.

The conflict between the ideas of meritocracy and the autonomy of family stems from a combination of facts about families and the structure of liberal societies.

Taking the conventional definition of merit as the product of effort and talent, we can state that, to some extent, merit-based success can be attributed to the individuals' effort and decisions in life. However, we also have to acknowledge that a substantial part of what is often labeled as talent is determined by a genetic component (Haworth et al. 2010; Plomin and Deary 2015; Plomin and Spinath 2002), early childhood experiences (Coneus, Laucht, and Reuß 2012; Heckman 2006, 2013) and social origin in general (Anger and Heineck 2010). Newer studies even show that the ability to exert effort has a similar way of being

determined early in life or at birth and might be beyond the influence of the individual (Anger and Schnitzlein 2017; Fletcher and Wolfe 2016). If both talent and effort are

empirically in large parts determined by factors outside the individual's control, this would question whether individual success as a product of talent and effort is in line with our general notion of merit-based success. In the word of James J. Heckman, the "accident of birth" (Heckman 2013) should not be a significant determining factor of life chances if we are to call a society equal in terms of opportunities for all individuals.

(15)

3 In my thesis, I will not go into the details of the normative discussion in this field but will contribute to the empirical study of conditions, which are used to evaluate to what degree equality of opportunity in societies can be claimed to exist. Even without the normative dimension, this is a fiercely contested area of research in sociology and economics. The political relevance stems from the fact that investigations in this area do not only hold significance for specific focused policy measures or regulations. The results in this field of study and my thesis aim at the heart of modern liberal societies as they pose a (necessary) challenge to the myth of equality of opportunity. A thorough sociological analysis needs to consider this broad perspective, even if any empirical application might always seem limited due to the strong contextualization inherent in such analysis.

My focus will lie on the role that the family plays in shaping individuals' life chances and their opportunities in life. I will show that family carries a tremendous amount of influence in different spheres of life, like career development, family formation, and the general life course of individuals. My guiding hypothesis is that the influence of the family of origin affects almost all areas of an individual’s life through different channels, each with their own set of mechanisms.

1.2 The research gap in the intergenerational transmission of inequality

When narrowing down on whether and to what degree families shape the life chances of their children, sociologists have often used single indicators of social class or parental

education. These variables are usually treated as proxies for systematic differences between families that are part of society's broader social structure and might afford children

differential access to resources and success in later life. Suppose a systematic association between the parents' social position and the social position of the children is found in statistical analyses. In that case, we interpret this as evidence for intergenerational

transmission of inequality and a potential flaw in the liberal order of society. The flaw lies in the fact that under ideal situations, the liberal view on society would state that we have perfect mobility, meaning that the individual's success is not related to the family of origin. Two well-established approaches within this larger area of research include sociological studies of class mobility (e.g. Breen and Jonsson 2005; Erikson and Goldthorpe 1992; Hertel and Groh-Samberg 2014) and economists' analyses of income or wage mobility (e.g.

(16)

4 Björklund and Jäntti 2011: Corak 2004; Couch and Dunn 1997; Lee and Solon 2009). While the former is traditionally based on mobility tables, the latter usually works with parent-child (typically father-son) correlations of income or wages.

While studies in class mobility are usually concerned with the persistence and fluidity of the importance of social origin for the destination class over historical time or in cross-country comparisons, income mobility studies usually try to identify causal mechanisms through which the parent-child correlation in income can be explained. Empirical results have shown that for some countries, like the U.K., intergenerational income mobility and

intergenerational class mobility do not align (Blanden et al. 2013). This discrepancy has opened the discussion about a so-called mobility paradox and the question of whether intergenerational class inequality and intergenerational income inequality relate to different social phenomena or if they instead are two indicators of one underlying inequality process (Breen, Mood, and Jonsson 2016).

One thing that has been consistently shown, though, is that there is no secular trend of declining inequality of opportunity throughout the OECD world (e.g. Breen and Müller 2020). It seems the intergenerational transmission of inequality is harder to eradicate than the liberal paradigm would suggest or wish for.

One problem that consistently comes up in studies of mobility is that they focus on only one dimension of the socio-economic position of the family. Unfortunately, the decision of which dimension to choose seems to be driven mostly by disciplinary traditions and less by

thoughtful theoretical considerations or empirical guidelines. The problem is that studies looking into the micro-mechanisms of intergenerational transmission of inequality cast doubt on the idea that one dimension alone can encompass the plethora of pathways through which parents bestow their children with advantages or settle them with burdens. In my thesis, I argue that it becomes particularly relevant to consider the multiplicity of the family of origin if one looks beyond the narrow scope of a single marker of success in

peoples’ lives but instead analyses different dimensions of inequality over the course of life. I therefore propose to use sibling similarities in outcomes as a much more general measure of the total influence of the family of origin (e.g. Björklund, Jäntti, and Lindquist 2009; Conley and Glauber 2007; Hauser and Wong 1989; Mazumder 2008; Solon et al. 1991), that goes beyond single or even multidimensional measures of social positions, previously used in the literature. I demonstrate that this approach allows for a broadening of the theoretical

(17)

5 view on intergenerational transmission of inequality and empirically has substantial

additional explanatory power over and above traditional measures of the social position of the family of origin.

My thesis brings together advances in theoretical discussions on what social origin means, how we should measure it, and through what mechanisms it becomes relevant for the life of children. Additionally, I expand the traditional analyses of single outcomes to cover different areas of labor market success as well as family formation, demonstrating the seemingly omnipresent relevance of our family of origin throughout the life course. This approach of using the sibling similarity as a measure of social origin both requires and allows for a more comprehensive notion of how the family structures the life chances of their children. It includes impacts of the neighborhood the family chooses to live in, the school that the children are sent to, and any other way in which parents influence their children. Taking this comprehensive notion of familial influence as well as our understanding that these empirical analyses can be of fundamental importance for how we judge the fairness and equity of our societal system, I develop a thesis that has four empirical studies at its core.

1.3 Sibling similarity in socio-economic status attainment

In recent years a research framework has emerged that measures sibling similarity in socio-economic status to assess the total effect of the family background on status attainment. The idea behind this framework is that siblings share specific family background

characteristics that make them similar in their later attainment. These shared family

influences can be, for example, parental resources, parenting strategies, and shared genetic endowments. In this way, the sibling similarity represents a compound measure of family background that entails both observable and unobservable aspects of family background, which is arguably the most comprehensive measure for the importance of the family of origin for status attainment. So far, there is no agreement in the literature on how much of the family effect is captured by the sibling similarity. Since the sibling similarity in a status outcome captures everything siblings share, it potentially also captures shared effects of the neighborhood, as well as influences that siblings have on each other (Mazumder 2008). Therefore, some authors argue that sibling similarity stands for the upper bound of the

(18)

6 family background effect or even overestimates the family background effect (e.g.

Beenstock 2008). If we do not follow the argument that neighborhood effects are part of the overall family effect since parents to a certain degree choose where they live and to what school they send their children, then having these effects also captured by the sibling similarity is problematic. However, some studies tried to disentangle the influence of the neighborhood from the sibling similarity and found these effects to be very small so that they are unlikely to bias the sibling similarity estimation (Bügelmayer and Schnitzlein 2018; Lindahl 2010).

On the other hand, the sibling similarity does not capture the differential treatment of parents. In the case where parents, for example, invest differently in their children, whether based on gender or other characteristics. These investments are no longer shared by siblings and thus not part of the sibling similarity. Nevertheless, these parental behaviors are also part of the influence the family background has on later status attainment. For this reason, Conley et al. (2007) argue that we should also look at differentiation processes within the family if we are interested in the importance of social background for status attainment. They point to the sex composition of the sibling group and the position in the birth order as possible characteristics by which parents differentiate their behavior. In this way, it can be argued that the sibling similarity represents the lower bound of the family background effect as it does not capture these child-specific family influences (Björklund and Jäntti 2012).

1.3.1 Sibling similarity for different dimensions of socio-economic status attainment

Research on sibling similarity is concerned with several different outcomes of status attainment. Most research in sociology is concentrated on educational attainment (e.g. Aizer, Stroud, and Buka 2012; Beenstock 2008; Björklund and Jäntti 2012; Conley and Glauber 2005; Grätz 2018; Hauser and Sewell 1986; Hauser, Sheridan, and Warren 1999; Hauser and Wong 1989; Lindahl 2010; Marks and Mooi-Reci 2016; Sieben and de Graaf 2001; Sieben, Huinink, and de Graaf 2001; Toka and Dronkers 1996; Wiborg and Hansen 2018), while the focus of economic research lies mostly on the similarity in earnings and income (e.g. Altonji and Dunn 2000; Beenstock 2008; Björklund et al. 2002; Björklund, Jäntti, and Lindquist 2009; Björklund, Lindahl, and Lindquist 2010; Björklund and Jäntti 2012; Conley and Glauber 2005; Mazumder 2008; Page and Solon 2003; Schnitzlein 2014; Sieben

(19)

7 2001; Wiborg and Hansen 2018). Among other outcomes that have been studied are

occupational status (e.g. Conley and Glauber 2005; Hauser et al. 1999; Toka and Dronkers 1996; Warren, Sheridan, and Hauser 2002), wealth (e.g. Conley and Glauber 2005), hours of work (Altonji and Dunn 2000), cognitive skills (e.g. Björklund and Jäntti 2012; Duncan, Boisjoly, and Harris 2001; Hauser et al. 1999) and non-cognitive skills (Anger and Schnitzlein 2017).

1.3.2 Results for different countries

Estimating sibling similarities poses high demands on data that are usually hard to satisfy. One needs to have not only information on the individual's status attainment but also on his or her siblings. This need for complex data is likely one reason why most studies use either survey data from the United States (e.g. PSID or Wisconsin Longitudinal Study) or

Scandinavian register data and why evidence for other countries is much less common. Compared to the United States, the similarity of earnings was found to be lower in

Scandinavian countries, leading to the conclusion that family factors are more important for socio-economic status attainment in the United States (Björklund et al. 2002). For other countries, the evidence is much less comprehensive. A recent study estimated the sibling similarity in the educational attainment in Australia and found a decline in the influence of overall family background over time (Marks and Mooi-Reci 2016). Another study looked at the similarity in educational attainment, wealth, and occupational prestige in Hungary, during times of Communism (Toka and Dronkers 1996). Based on a unique dataset for the Netherlands, Knigge et al. (2014a; 2014b) calculated brothers' correlations in occupational status for the 19th century, the time of modernization, of the size of 0.52, which is higher

than comparable measures for the Netherlands today. They also found a decrease in the similarity, leading them to conclude that family background became less critical for status attainment, although processes of modernization did not cause this development. In a structural equation modeling framework, deGraaf and Huinink (1992) look at the similarity of siblings in West Germany in educational attainment and occupational status. Another study by Sieben, Huinink, and de Graaf (2001), which also uses structural equation models, has looked at the sibling resemblance in educational attainment in Germany and found a weaker similarity in the former German Democratic Republic than in the former Federal

(20)

8 Republic of Germany. Based on data from the German Socioeconomic Panel Study, a recent study focuses on the total family effect on earlier life outcomes of school grades and

cognitive skills plus various health measures (Bügelmayer and Schnitzlein 2018). With the same dataset, Grätz (2018) studies the sibling similarity in educational attainment and sets the inequality between families into context with within- family inequality by birth order and maternal age at birth.

Studies that explicitly compare the strength of the total family background across different countries are also not common. They usually find comparable sizes of the sibling similarity, although the estimates for the U.S. are usually higher than the estimates for Scandinavian countries, with Germany falling in-between. As Björklund et al. (2002) show, the similarity in earnings is smaller in Scandinavian countries (ranging from 0.14-0.26) than in the U.S. (0.40), leading to the conclusion that the family is overall more important for determining earnings in the United States. Comparing the total family effect on educational attainment and occupational status in countries with different levels of modernization has shown a decline in the effect on the occupational status with increasing levels of modernization (Sieben and de Graaf 2001). However, the decline in educational attainment does not seem to be significant, meaning the modernization hypothesis is only partly supported. Another study looked at the sibling similarities of earnings in Germany in comparison to other countries. It has shown a similarity for brothers, which is similar in strength to the United States, but higher than in Denmark. For sisters, the similarity is higher than in both the U.S. and Denmark (Schnitzlein 2014). Most recently, a study focusing on a comparison of the sibling similarity in Nordic countries, Germany, the U.K., and the U.S. in educational attainment, grades, and cognitive ability has shown only small cross-country differences (Grätz et al. 2019).

1.3.3 Results for different social groups

Only some studies look at the sibling similarity for different social groups. Conley and Glauber (2008) show that sibling similarity in educational attainment, earnings, and family income in the United States varies with the education, age, and family status of the mother, as well as with the number of siblings in a family. They also find that siblings from more advantaged families have a higher similarity than siblings from disadvantaged families. This

(21)

9 variation in sibling similarity by social class background means that for siblings from

disadvantaged families, the family background is less relevant for their socio-economic attainment than for siblings from more advantaged backgrounds. However, in a second study, Conley, Pfeiffer, and Velez (2007) find that siblings from disadvantaged families are more like each other in their performance in achievement tests than siblings from

advantaged families, which in turn means that the family background has a more substantial influence for disadvantaged siblings. Thus, the importance of the family background for different aspects of attainment seems to differ concerning the social group but also concerning the specific outcome that is studied. In addition to social background, there is evidence that the similarity of siblings differs by race, as African-Americans are reported to have a lower similarity in socio-economic outcomes than white Americans (Conley and Glauber 2007). However, for Germany, Grätz (2018) does not find any substantial differences in educational achievement by parental socio-economic background.

Furthermore, Schnitzlein (2012) finds no relevant difference in the similarity in permanent earnings between natives and different immigrant groups in Denmark. All in all, the evidence on differences in sibling similarity by population subgroups appears to be inconclusive, and further systematic and theory-driven research is needed.

1.3.4 Sibling similarity over the life course

The variation of the similarity over the life course is also a topic where the evidence is inconsistent. Two studies by Conley and Glauber (2005, 2007) use PSID data and find no evidence for changing sibling similarity in education and occupational status, but rather the converging similarity in income and earnings over the life course. In another study, Hauser et al. (1999) used Wisconsin Longitudinal Study data and found a stable sibling similarity in occupational standing. In contrast to this finding, Warren et al. (2002) find an increase in the sibling similarity in occupational standing using the same data, but with a more sophisticated modeling strategy. Also, Erola (2012) finds quite a stable similarity in education (measured in levels and years) and in occupational status (EGP and ISEI), using Finnish census panel data. He measures each outcome twice between the ages of 20-29, 25-34, 30-39, and 35-44. In a different approach, Erola et al. (2016) research whether observable parental background characteristics, like paternal and maternal education, occupational status and income,

(22)

10 measured at different points during the childhood of siblings, can explain different parts of the sibling similarity in adult occupational status outcomes, which were measured at age 25-29 and 30-34. They find no discernible life course differences in the influence of observable parental characteristics, leading them to conclude that not repeatedly measuring parental background characteristics does not significantly bias results.

All in all, the existing evidence does not paint a consistent picture, and no clear pattern of divergence or convergence of sibling similarity in socio-economic status emerges. The evidence seems to either point to a stable influence of parental background or siblings becoming more similar over the life course, which can be interpreted as increasing importance in family background.

In conclusion, research on sibling similarity has covered a variety of different socio-economic status measures, but most of the studies are based on data from either the U.S. or Nordic countries. While most studies find higher similarities in educational outcomes than in labor market outcomes, results for differences between social groups and for the development of the similarity over the life course are less conclusive. Here further research is needed. There also seems to be a division in proposals for further research. On the one hand, there is a need to study further the decomposition and the underlying mechanisms of the sibling similarity; on the other hand, one could argue that the sibling similarity should best be used as a compound measure of family background that has the advantage of capturing

unobservable influences.

1.4 Methods – A sibling-based approach to the estimation of intergenerational inequality

1.4.1 Measures and methods

There exist two different methodological approaches in the literature to estimate the sibling similarity. The most common approach defines the sibling similarity as the

intra-class-correlation (ICC) in a multi-level framework. The ICC gives us the relative explanatory power of the family and sibling level. The sibling correlation thus is the relative size of the family variance to the sum of family variance and sibling variance (Mazumder 2008). Compared to traditionally used parent-child correlations, the sibling correlation has the advantage of capturing the intergenerational transmission process more comprehensively. The sibling

(23)

11 correlation in income, for example, is equivalent to the parent-child correlation in income squared plus other shared influences on income between siblings, which are uncorrelated to the parental income (Björklund at al. 2010).

In an approach developed by Solon et al. (1991), the individual time-specific variation is also estimated by taking repeated observations of an individual into the model. If the time-specific variance component is excluded from the calculation of the ICC, this approach yields a time-stable measure of sibling similarity within the period of observation. Solon et al. (1991) argued that a time-stable measure of sibling similarity is preferable to one based on only single individual observations since measurement error can be reduced using this approach. Using just single observations would thus underestimate the sibling similarity. An alternative approach estimates sibling similarity using a Structural Equation Modelling (SEM) framework. This design was first proposed by Hauser et al. (1989), who used SEM for a series of papers based on the Wisconsin Longitudinal Study. It is also used by Sieben et al. (2001) but seems overall less common. Since in the SEM framework, the correlation is established between two siblings and then weighted for the number of overall siblings, the size of the correlation is directly comparable to the ICC, estimated only for sibling pairs. Inge Sieben (2001) wrote an overview of the advantages and disadvantages of the multi-level framework compared to the SEM framework. She sees the advantage of the SEM approach mainly in the possibility to estimate path models and test for significant trends in effects while the multi-level model makes it easier to introduce context effects and analyze categorical outcomes. In this thesis, I decided to use multi-level modeling as my analytical strategy for methodological reasons that are further discussed in section 1.4.6. Also, this is the methodological approach that has been utilized in most recent studies on sibling

similarity in sociology and economics and thus makes my results more easily comparable to the broader discussion of sibling similarity in the literature.

(24)

12

1.4.2 Siblings and Intra-class correlation

A formal regression equation for the multi-level model without predictors, which is used in this thesis, is defined as:

𝑦𝑓𝑠𝑡 = 𝛼 + 𝑢𝑓+ 𝑣𝑓𝑠+ 𝑒𝑓𝑠𝑡 (1)

In equation one, the index 𝑓 stands for family, 𝑠 for a sibling, and 𝑡 for time. The data structure is such that time points of observations are nested within individual siblings, and those siblings are again nested within the family. It is assumed that the random components 𝑢𝑓, 𝑣𝑓𝑠𝑎𝑛𝑑 𝑒𝑓𝑠𝑡 are uncorrelated. The constant 𝛼 is the overall average of the outcome in the dataset. The total variance of the dependent variable can be decomposed into three parts on the three levels:

𝑉𝑎𝑟(𝑦𝑓𝑠𝑡) = 𝜎𝑓2+ 𝜎𝑠2+ 𝜎

𝑡2 (2)

From this follows that the total variance is the sum of the variance between families, the differences between siblings, and the differences between different points of measurement in time for each sibling. For my thesis, the time-specific differences are not of interest. Unless otherwise specified, I aim to measure a time-stable outcome. Differences at specific time points are treated as measurement error, as deviations from this time stable measure. Excluding the time-specific variance component is an approach pioneered in the seminal paper by Solon et al. (1991). They demonstrate that the degree of intergenerational transmission of inequality is massively underestimated if only specific points in time (with high degrees of measurement error) are used for the calculation. Therefore, the so-called intra-class correlation (ICC) 2, which is used in many multi-level applications, is calculated

based only on the variance components for the siblings and the family.

2 The ICC is also sometimes called the variance partitioning coefficient (Goldstein, Browne,

and Rasbash 2002). It can be argued that this is technically speaking the more accurate term, but in this thesis, I will stick to the usage of the term ICC to be in line with the applied multi-level literature.

(25)

13 In the context of sibling models, the ICC (𝜌) is also referred to as sibling correlation. It is thus the relative size of the family variance to the sum of family variance and sibling variance, excluding the time-specific variance (see equation 3). In this sense, it represents time stable similarities between siblings or time stable differences between families. The two

interpretations are statistically speaking equivalent in this context, and the interpretation of the ICC depends on the research question.

𝜌 = 𝜎𝑓 2

𝜎𝑓2+𝜎𝑠2 (3)

Time stable always implies time stable within the period of observation. For example, if we estimate sibling similarity, or family differences, in income between year one and three after labor market entry, we take the average of repeated measurements of income per sibling as a measure of time stable income. Measurement error is reduced by this method because time-specific deviations from the time-stable income are excluded.

The general approach of comparing the within (families) and between (families) variance is common throughout the literature. The advantage of taking the proportion of the variance explained between families compared to the total variance lies in the fact that the ICC is independent of the dependent variable's scale, which allows for comparisons across different types of dependent variables. It is not a variable but a ratio of two estimates (variance estimates on both levels) from one model. Roughly speaking, the ICC would be approximately equivalent to the 𝑅2 we would get in a linear regression if we included a dummy variable for each family.

The fact that the ICC is bounded by 0 and 1 yields ideal-typical situations that can help to gauge the magnitude of the sibling similarity. An ICC of 1 in labor income would mean that all the variation in income is due to differences between families. Within the family, all siblings would have the same income, and thus there is no variation within families (or between siblings). This hypothetical scenario would mean that the family of origin wholly determines labor income in such a scenario. This scenario thus constitutes the most extreme case of transmission of inequality. An ICC of 0 in labor income, on the other hand, would mean that all the variation in income is due to differences between siblings, and families would, on average, have the same income. In this scenario, we would say that the family of

(26)

14 origin does not influence income differences. Only the siblings' idiosyncratic characteristics – those they do not share with their other siblings – determine the labor market income. As the ICC is an estimated quantity, I report confidence intervals for the ICC for the various status outcomes. The confidence intervals are based on a standard error for the ICC, which Stata calculates via the delta-method.3 All models are estimated using restricted maximum

likelihood (REML). In contrast to conventional multi-level models, this approach considers the degrees of freedom that are used to estimate the coefficients of the model when estimating the variance of the error terms. In standard multi-level models, not taking this into account might often lead to an underestimation of the variance components.

1.4.3 Explaining the ICC through observed variables

The ICC in the multi-level sibling models gives an overall estimate of the total familial influence on the outcome under study. It constitutes the sum of all influences on the outcome which are shared by the siblings. One of the research questions that I will address in this thesis is how much of this overall familial influence can be explained by a set of observed variables. For this purpose, I will choose measurements of social origin, which are traditionally used in studies of intergenerational transmission of inequality. If I introduce these variables into equation 5, I get estimates of the different variance components that are conditional on the set of observed variables. The change can consequently be

interpreted as a part of intergenerational transmission or sibling similarity, which can be explained by these variables.

𝑦𝑓𝑠𝑡 = 𝛼 + 𝑋𝑓𝑠𝑡𝛽 + 𝑢𝑓̈ + 𝑣𝑓𝑠̈ + 𝑒𝑓𝑠𝑡̈ (4)

𝜌̈|𝑋 = 𝜎𝑓 2̈ 𝜎𝑓2̈ +𝜎𝑠2̈

(5)

For example, suppose 𝑋 contains parental education and occupation. In that case, the difference between the ICC before and after introducing the variables can be interpreted as

3 I use the nlcom command, which can be used to approximate standard errors for

(27)

15 the part of the overall transmission of inequality in this outcome, which is due to differences in parental education and occupation. If these traditional variables capture most of the transmission process, we expect the conditional ICC to be close to zero. If they play almost no role, the conditional ICC should be like the unconditional ICC.

1.4.4 Number of siblings and ICC

Another criticism that might come up is that the degree of within variance depends on the number of siblings. The intuitive idea would be that more siblings generate more diversity and, therefore, more variation, and consequently, the ICC would become lower the higher the number of siblings in the sample. However, this idea conflates every-day notions of diversity with the statistical notion of variance. We can reject this idea just by thinking about what this would mean for the estimation of the variance of any variable from a sample. If it is a random sample, the sample variance is an unbiased estimator of the true population variance. If the claim of more units leads to more variance were true, the sample variance would be increasing the larger the sample is. This scenario is inconsistent with the idea of an unbiased estimator. When we consider how common sample variances are calculated, we can see that the number of observations is considered:

𝜎𝑦2 = 1 𝑛−1∑ (𝑦𝑖 − 𝑦̅) 2 𝑁 𝑖=1 (6)

The same is true for the within variance in with 𝑦̅ represents the family average, and the other values the siblings' values. Larger families do not mechanically produce lager within variances. If we find that larger families produce larger within variance, this would be an empirical finding and must reflect a substantive underlying (social) process.

1.4.5 Comparing multi-level to SEM or mobility tables

Many studies of intergenerational transmission of inequality and social mobility have relied upon (modified) mobility tables, which essentially cross-tabulate the class of origin and class of destination. The cross-tabulation can be modeled in a log-linear model of table cell

(28)

16 corresponding model fits are compared to assess the hypotheses which best describe

patterns of social reproduction from one generation to the next.

These models investigate the structure of dependencies of one outcome, class of

destination, to one parental characteristic, class of origin. In my approach, I also focus on one outcome. However, using the sibling-ICC approach, this outcome is not related to one specific parental characteristic, but the sum of all influence from the family, which is shared by the siblings. The implication here is that factors other than the class of origin, e.g.

education or the family structure, also drive the degree of inequality transmission. Further, a mobility table might look different for the first or the second child and can be applied to only children as well. The ICC is an estimate that is applied equally to all children within families and excludes by design only-children. Therefore, the ICC approach is not necessarily superior or inferior to mobility tables but carries the advantage of being an overall measure of the influence of the family of origin, which is not restricted to one parental characteristic. At the same time, only-children cannot be analyzed using this approach.

It is also possible to estimate the sibling similarity within a structural equation modeling framework (SEM). In this case, each sibling would be modeled by a separate outcome variable in a separate equation. This approach usually requires the artificial reduction of the number of siblings per sibling pair to two. The correlation of the two outcomes (potentially conditional on a set of predictor variables) would then be interpreted as the sibling

correlation and is statistically equivalent to the ICC measure from a model that is estimated on the same data, meaning only two siblings per family. The SEM approach shows no advantages for my research questions and has the limitation that it cannot easily make use of more than two siblings per family, which leads to a lot of unused information.

Therefore, my results can be compared to previous studies choosing the SEM approach, but the multi-level approach is strongly preferable for my research agenda.

1.5 Structure of the thesis

The thesis is structured in the following way. In the previous section, I gave a general overview of the relevant literature on sibling similarity in socio-economic outcomes and a generalized overview of the methodology of sibling similarity studies used in this thesis. This section set out the research gap that the four studies of this thesis aim to fill and defined

(29)

17 how sibling similarity, as a comprehensive measure of social origin, allows me to address key challenges in the current literature.

Following this, in chapter two, I present my study on the direct effects of social origin on labor market success throughout the early career. This chapter explores the relationship between social origin, education, and labor market destination using sibling models. I argue that one of the shortcomings of existing studies on the intergenerational transmission of inequality lies in their operationalization of social origin, which is usually based on single indicators, like parental social class or parental education. Therefore, I propose using an approach in which the sibling similarity represents a compound measure of social origin that entails both observed and unobserved aspects of family background. I aim to answer three research questions. First, is there a direct effect of social origin on labor market attainment which is not mediated by education? Second, can traditional measures of social origin account for the direct effect of social origin estimated by sibling models? Third, does the influence of social origin on labor market attainment change over the early labor market career? I estimate sibling correlations for Germany using SOEP data. I consider three measures of labor market attainment, namely labor income, occupational status, and position in the job hierarchy, each measured at different points in time over the career to capture the multidimensional nature of labor market attainment. Results show that the mediating effect of education explains a significant amount of the sibling similarity for all three status outcomes. However, a substantial amount of sibling similarity is not explained by educational attainment. Moreover, while observed social origin measures account for a large part of the sibling similarity, a small part is left unexplained. Besides, there is no clear pattern regarding the divergence or convergence of the level of sibling similarity in labor market attainment over the early career.

The third chapter addresses differences in the intergenerational transmission of inequality in labor market outcomes between different social origin groups. In this study, we estimate whether and how sibling similarity in labor market outcomes in Germany varies by economic background. We try to answer the question: is the sibling similarity in socio-economic outcomes stronger in upper-class than in lower-class families? We use data on siblings from the SOEP and make two contributions to the literature. First, we propose not only to look at the intra-class correlation (ICC) as a measure of sibling similarity but also to take both the within- and between-family dispersion into account individually. Second, we

(30)

18 consider different socio-economic outcomes and show how sibling similarity varies by social origin, depending on the type of socio-economic outcome considered. Socio-economic status is measured as occupational status, labor market earnings, and job hierarchy position between the ages of 18 and 45. Results show that sibling similarity in occupational status is lower in the lowest social class, while there is no clear trend for labor market income and job hierarchy position. Different strategies for the reproduction of social privilege seem to be simultaneously at play and must be investigated in the future by a sharper focus on measuring the mechanisms.

The fourth chapter leaves the realm of transmission of labor market success and focuses on union formation instead. Being in a committed relationship, cohabiting with a partner, or getting married are essential events in the transition to adulthood. Recent studies have shown that family background characteristics play a role in determining the timing of these events. In this study, I investigate the similarity of siblings in the timing of first living-apart-together union, first cohabitation, and first marriage. In a second step, I assess whether parental socio-economic resources and family formation behavior and individual

characteristics of the siblings can account for the similarity between siblings. The results indicate that siblings are more similar in their timing of first marriage than in their timing of first LAT union or cohabitation. Sibling similarity in union timing is only minimally reduced through the inclusion of observed parental and individual characteristics. The observed parental characteristics do predict the timing of union but cannot explain the similarity between siblings.

The fifth chapter retains the focus on siblings, but in contrast to the first three studies of the thesis, which focus on inequality between families, in this study, I am researching how inequality is generated within the families. I will investigate how within-family dynamics, namely the gender composition of the sibling group, influence inequalities regarding occupational attainment. To this date, occupational gender segregation and a substantial gender pay gap are defining characteristics of labor markets in Western societies. While it has been established that access to occupations is gendered, it is less clear whether there are selection effects into different types of occupations by the gender of one’s sibling(s). I differentiate between horizontal occupational stratification and vertical occupational stratification to investigate how the gender composition of siblings in interaction with the birth order influences these two labor market outcomes at the beginning of the career. I

(31)

19 look at the differences in occupational choice and attainment between those who have same-gender sibling(s) versus those with opposite-gender sibling(s). The leading question of this study is: Does having a sibling of the opposite gender increase the chances of choosing a gender-atypical occupation at the entry of the labor market career? I derive a set of

hypotheses about the impact of sibling gender composition and its consequences from an extension of established theories in the field. The empirical results indicate that how these factors matter is more complicated than these theories predict. It seems that having a younger sibling of the opposite gender is detrimental to occupational success when

compared to having a younger sibling of the same gender. While within-family dynamics are an important aspect in the way the family shapes the life course and labor market

trajectories for individuals, more research is needed to investigate what theoretical mechanism can explain this finding.

The last chapter concludes by summarizing the results of this thesis, interpreting them in light of the existing literature and the theoretical developments that can be made from them and draws some more general conclusions about the way we see the role of the family in the process of generating and maintaining inequality within liberal societies. I finish by giving an outlook on promising future paths of research.

(32)

20 1.6 References

Aizer, Anna, Laura Stroud, and Stephen Buka. 2012. Maternal Stress and Child Outcomes: Evidence from Siblings. Working Paper. 18422. National Bureau of Economic Research. Altonji, Joseph G., and Thomas A. Dunn. 2000. "An Intergenerational Model of Wages,

Hours, and Earnings." The Journal of Human Resources 35(2):221–58.

Anger, Silke, and Guido Heineck. 2010. "Do Smart Parents Raise Smart Children? The

Intergenerational Transmission of Cognitive Abilities." Journal of Population Economics 23(3):1255–82.

Anger, Silke, and Daniel D. Schnitzlein. 2017. "Cognitive Skills, Non-Cognitive Skills, and Family Background: Evidence from Sibling Correlations." Journal of Population Economics 30(2):591–620.

Beenstock, Michael. 2008. "Deconstructing the Sibling Correlation: How Families Increase Inequality." Journal of Family and Economic Issues 29(3):325–45.

Björklund, Anders, Tor Eriksson, Markus Jäntti, Oddbjörn Raaum, and Eva Österbacka. 2002. "Brother Correlations in Earnings in Denmark, Finland, Norway and Sweden Compared to the United States." Journal of Population Economics 15(4):757–72.

Björklund, Anders, and Markus Jäntti. 2011. "Intergenerational Income Mobility and the Role of Family Background." The Oxford Handbook of Economic Inequality.

Björklund, Anders, and Markus Jäntti. 2012. "How Important Is Family Background for Labor-Economic Outcomes?" Labour Labor-Economics 19(4):465–74.

Björklund, Anders, Markus Jäntti, and Matthew J. Lindquist. 2009. "Family Background and Income during the Rise of the Welfare State: Brother Correlations in Income for Swedish Men Born 1932–1968." Journal of Public Economics 93(5–6):671–80.

Björklund, Anders, Lena Lindahl, and Matthew J. Lindquist. 2010. "What More Than Parental Income, Education and Occupation? An Exploration of What Swedish Siblings Get from Their Parents." The B.E. Journal of Economic Analysis & Policy 10(1).

Blanden, Jo, Paul Gregg, and Lindsey Macmillan. 2013. “Intergenerational Persistence in Income and Social Class: The Effect of within-Group Inequality.” Journal of the Royal Statistical Society: Series A (Statistics in Society) 176(2):541–63.

Boudon, Raymond. 1974. "Education, Opportunity, and Social Inequality: Changing Prospects in Western Society."

Breen, Richard, ed. 2004. Social Mobility in Europe. Oxford University Press.

Breen, Richard, and Jan O. Jonsson. 2005. "Inequality of Opportunity in Comparative Perspective: Recent Research on Educational Attainment and Social Mobility." Annual Review of Sociology 31:223–43.

Breen, Richard, Carina Mood, and Jan O. Jonsson. 2016. "How Much Scope for a Mobility Paradox? The Relationship between Social and Income Mobility in Sweden."

Sociological Science 3:39–60.

Breen, Richard, and Walter Müller, eds. 2020. Education and Intergenerational Social Mobility in Europe and the United States. Stanford University Press.

Bügelmayer, Elisabeth, and Daniel D. Schnitzlein. 2018. "Is It the Family or the

Neighborhood? Evidence from Sibling and Neighbor Correlations in Youth Education and Health." The Journal of Economic Inequality 16(3):369–88.

Coneus, Katja, Manfred Laucht, and Karsten Reuß. 2012. "The Role of Parental Investments for Cognitive and Noncognitive Skill Formation—Evidence for the First 11 Years of Life." Economics & Human Biology 10(2):189–209.

(33)

21 Conley, Dalton, and Rebecca Glauber. 2005. Sibling Similarity and Difference in

Socioeconomic Status: Life Course and Family Resource Effects. Working Paper. 11320. National Bureau of Economic Research.

Conley, Dalton, and Rebecca Glauber. 2007. "Family Background, Race, and Labor Market Inequality." Annals of the American Academy of Political and Social Science 609:134– 52.

Conley, Dalton, and Rebecca Glauber. 2008. "All in the Family?: Family Composition, Resources, and Sibling Similarity in Socioeconomic Status." Research in Social Stratification and Mobility 26(4):297–306.

Conley, Dalton, Kathryn M. Pfeiffer, and Melissa Velez. 2007. "Explaining Sibling Differences in Achievement and Behavioral Outcomes: The Importance of within- and between-Family Factors." Social Science Research 36(3):1087–1104.

Corak, Miles. 2004. Generational Income Mobility in North America and Europe. Cambridge University Press.

Couch, Kenneth A., and Thomas A. Dunn. 1997. "Intergenerational Correlations in Labor Market Status: A Comparison of the United States and Germany." The Journal of Human Resources 32(1):210–32.

De Graaf, Paul M., and Johannes J. Huinink. 1992. "Trends in Measured and Unmeasured Effects of Family Background on Educational Attainment and Occupational Status in the Federal Republic of Germany." Social Science Research 21(1):84–112.

Duncan, Greg J., Johanne Boisjoly, and Kathleen Mullan Harris. 2001. "Sibling, Peer,

Neighbor, and Schoolmate Correlations as Indicators of the Importance of Context for Adolescent Development." Demography 38(3):437–47.

Erikson, Robert, and John H. Goldthorpe. 1992. The Constant Flux: A Study of Class Mobility in Industrial Societies. Oxford University Press.

Ermisch, John, Markus Jäntti, and Timothy M. Smeeding. 2012. From Parents to Children: The Intergenerational Transmission of Advantage. Russell Sage Foundation.

Erola, Jani. 2012. The Life Course Variation of Sibling Correlations According to Class and Education. SSRN Scholarly Paper. ID 2133753. Rochester, NY: Social Science Research Network.

Erola, Jani, Sanni Jalonen, and Hannu Lehti. 2016. "Parental Education, Class and Income over Early Life Course and Children's Achievement." Research in Social Stratification and Mobility 44:33–43.

Fletcher, Jason M., and Barbara Wolfe. 2016. "The Importance of Family Income in the Formation and Evolution of Non-Cognitive Skills in Childhood." Economics of Education Review 54:143–54.

Goldstein, Harvey, William Browne, and Jon Rasbash. 2002. "Partitioning Variation in Multi-level Models." Understanding Statistics 1(4):223–31.

Grätz, Michael. 2018. "Competition in the Family: Inequality between Siblings and the Intergenerational Transmission of Educational Advantage." Sociological Science 5:246– 69.

Grätz, Michael, Kieron Barclay, Øyvind N. Wiborg, Torkild H. Lyngstad, Aleksi Karhula, Jani Erola, Patrick Präg, Thomas Laidley, and Dalton Conley. 2019. "Universal Family Background Effects on Education Across and Within Societies." MPIDR Working Paper WP 2019-007.

(34)

22 Hauser, Robert M., and William H. Sewell. 1986. "Family Effects in Simple Models of

Education, Occupational Status, and Earnings: Findings from the Wisconsin and Kalamazoo Studies." Journal of Labor Economics 4(3):83–115.

Hauser, Robert M., Jennifer T. Sheridan, and John Robert Warren. 1999. "Socioeconomic Achievements of Siblings in the Life Course New Findings from the Wisconsin Longitudinal Study." Research on Aging 21(2):338–78.

Hauser, Robert M., and Raymond Sin-Kwok Wong. 1989. "Sibling Resemblance and Intersibling Effects in Educational Attainment." Sociology of Education 62(3):149–71. Haworth, C. M. A., M. J. Wright, M. Luciano, N. G. Martin, E. J. C. de Geus, C. E. M. van

Beijsterveldt, M. Bartels, D. Posthuma, D. I. Boomsma, O. S. P. Davis, Y. Kovas, R. P. Corley, J. C. DeFries, J. K. Hewitt, R. K. Olson, S. A. Rhea, S. J. Wadsworth, W. G. Iacono, M. McGue, L. A. Thompson, S. A. Hart, S. A. Petrill, D. Lubinski, and R. Plomin. 2010. "The Heritability of General Cognitive Ability Increases Linearly from Childhood to Young Adulthood." Molecular Psychiatry 15(11):1112–20.

Heckman, James J. 2006. "Skill Formation and the Economics of Investing in Disadvantaged Children." Science 312(5782):1900–1902.

Heckman, James J. 2013. Giving Kids a Fair Chance. MIT Press.

Hertel, Florian R., and Olaf Groh-Samberg. 2014. "Class Mobility across Three Generations in the U.S. and Germany." Research in Social Stratification and Mobility 35:35–52.

Jackson, Michelle. 2007. "How Far Merit Selection? Social Stratification and the Labour Market." The British Journal of Sociology 58(3):367–90.

Knigge, Antonie, Marco H. D. van Leeuwen, and Ineke Maas. 2014a. "Sources of Sibling (Dis)Similarity: Total Family Impact on Status Variation in The Netherlands in the Nineteenth Century." AJS; American Journal of Sociology 120(3):908–48.

Knigge, Antonie, Ineke Maas, Marco H. D. van Leeuwen, and Kees Mandemakers. 2014b. "Status Attainment of Siblings during Modernization." American Sociological Review 79(3):549–74.

Lee, Chul-In, and Gary Solon. 2009. "Trends in Intergenerational Income Mobility." The Review of Economics and Statistics 91(4):766–72.

Lindahl, Lena. 2010. "A Comparison of Family and Neighborhood Effects on Grades, Test Scores, Educational Attainment and Income—Evidence from Sweden." The Journal of Economic Inequality 9(2):207–26.

Link, Bruce G., and Jo Phelan. 1995. "Social Conditions as Fundamental Causes of Disease." Journal of Health and Social Behavior 35:80–94.

Marks, Gary N., and Irma Mooi-Reci. 2016. "The Declining Influence of Family Background on Educational Attainment in Australia: The Role of Measured and Unmeasured

Influences." Social Science Research 55:171–85.

Mazumder, Bhashkar. 2008. "Sibling Similarities and Economic Inequality in the U.S." Journal of Population Economics 21(3):685–701.

Page, Marianne E., and Gary Solon. 2003. "Correlations between Sisters and Neighbouring Girls in Their Subsequent Income as Adults." Journal of Applied Econometrics

18(5):545–62.

Phillips, Anne. 2004. "Defending Equality of Outcome." Journal of Political Philosophy 12(1):1–19.

Plomin, R., and I. J. Deary. 2015. "Genetics and Intelligence Differences: Five Special Findings." Molecular Psychiatry 20(1):98–108.

Riferimenti

Documenti correlati

The main results in this section are the characterizations of the length of sets with countably many connected components (and in particular of continua) given in Theorem 2.5

31: Also at Warsaw University of Technology, Institute of Electronic Systems, Warsaw, Poland 32: Also at Institute for Nuclear Research, Moscow, Russia.. JHEP11(2018)152 33: Now

This new Topical Collection of the Journal of Nephrology on rare and complex kidney diseases aims to provide an updated overview on the incredible scope of research of

Las excavaciones de las tumbas de Las Lastras de San José (Albalate del Arzobispo) y de Planas de Esponera (Alcañiz) han proporcionado restos de jarritas de este tipo de

[r]

Statistically and clinically significant improvement was found in the parameters concerning comfort in production (males, p<0.001), sonorousness, vocal clarity

Our observations suggest that when pollination is not prevented stenospermocarpy may be responsible for the seedless phenotype of the Sangiovese somatic variant and is