The Geography of Intergenerational Mobility in Europe

The Geography of Intergenerational Mobility in Europe Olivia Granström, Per Engzell Document type Post-print: This manuscript has passed peer review and been accepted for publication by a journal. It may include final edits by the author(s) but no editing or formatting by the publisher. Funding information This research was funded by the European Research Council, Grant Agreement No. 101165962 Markets and Mobility: How Employers Structure Economic Opportunity. Suggested citation Granström, Olivia and Per Engzell. (2025). The Geography of Intergenerational Mobility in Europe. European Societies, forthcoming. Date of record October 10, 2025. Terms of use This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International license: https://creativecommons.org/licenses/by-nc-nd/4.0/ The Geography of Intergenerational Mobility in Europe Olivia Granström Stockholm University Per Engzell University College London October 10, 2025 Abstract How do opportunities for intergenerational mobility depend on where you live? We address this question using European Social Survey data, studying the association between parents’ and children’s occupation, and how it varies by region of residence. Absolute mobility, largely driven by shifts in occupational structure, differs from relative mobility, which reflects the extent to which opportunities are equal across social origins. Capital regions emerge as hubs of absolute, but not necessarily relative, mobility. Absolute mobility correlates with human capital, labor market, demographic, and spatial factors. In contrast, relative mobility is primarily shaped by economic disparities between social classes. Greater inequality entails less mobility, even comparing different places within a country. 1 Introduction Intergenerational mobility reflects the extent to which economic status persists across generations. Parents with greater skills and resources are better placed to invest in their children, and counteracting this imbalance is a key objective for social policy in liberal welfare states (Corak, 2013). It is difficult to determine the optimal level of mobility in society: most people probably hold that parents should be allowed some influence over their children. Nevertheless, comparing countries and regions can help identify places where opportunities are more or less equal (Breen and Jonsson, 2005). Beyond its link to opportunity, mobility may be important for outcomes such as social cohesion, democratic representation, or the efficient allocation of talent. Following a tradition that views the social division of labor as the backbone of stratification, we focus on occupational mobility as the basis for our analysis (Ganzeboom et al., 1991). Previous comparative work in this area has two key limitations. First, it has focused on country-level differences, attributing variation to national policies such as tax regimes and education systems. However, recent research suggests that opportunity varies almost as much within countries as between them (Chetty et al., 2014; DiPrete, 2020; Stuhler, 2018). Secondly, scholars have sought to locate the 1 determinants of mobility in childhood, neglecting how the labor market that people face may structure their opportunities (Engzell and Mood, 2023; Engzell and Wilmers, 2025). We advance the literature on both fronts, by studying how occupational mobility varies within and between European countries depending on the local labor market in which people reside. Earlier research on intergenerational mobility often measures contextual mechanisms during an individual’s childhood to capture childhood exposures that are argued to be the root causes of mobility. In this study, we map variation by place of adult residence to better understand how the labor market faced as adults may impact on mobility chances, as different places provide different economic opportunities. Specifically, access to jobs (Rothstein, 2019) and consequently migration to places where jobs are available (Buscha et al., 2021; Sprung-Keyser et al., 2022) have been associated with better chances of upward mobility. In Europe, different regions face dramatically different growth profiles, with some struggling to uphold and improve prosperity for their residents (Diemer et al., 2022; Morris and Oesch, 2025; Vigna, 2023). The link between social origin and occupational destinations may be strengthened when labor market prospects deteriorate, as young people become more reliant on parents’ social capital to find work (Morris, 2023; Zwysen, 2016). We draw on a large, harmonized cross-national dataset, pooling the first ten rounds (2002–2020) of the European Social Survey (ESS). Recent work on mobility disparities within countries has mostly analyzed administrative population data (Acciari et al., 2022; Brandén, 2019; Carneiro et al., 2019; Chetty et al., 2014; Chuard and Grassi, 2020; Eriksen and Munk, 2020; Güell et al., 2018; Heidrich, 2017; Kenedi and Sirugue, 2023). Such work can provide a detailed portrait of a given country, revealing how outcomes vary by region, city, neighborhood, or even block. Yet, the fact that administrative data are collected for different purposes than research makes them less comparable across countries. There has been limited work bridging the literature on within-country geographic disparities with the crosscountry research that came before it. Our use of an established multi-purpose survey allows us to do just that. Moreover, we explore a range of potential determinants of mobility and make our results available for other researchers to build on. Occupations can be measured in several ways. We focus on occupational rank in our main results, and follow up with results using social class. The rank approach assigns a score based on the typical education and income of an occupation, and then ranks it within each country’s distribution. Doing so lets us conceive of the social structure as a ladder, where each person has a proportion of the population above and below her. We further distinguish between absolute upward and relative mobility. Upward mobility reflects how far a person born at the bottom can expect to rise. Relative mobility refers to the correlation between parent and child occupation. The two are distinct. If a place undergoes occupational upgrading relative to the country mean, there will be net upward mobility. However, if privileged children benefit as much or more from this upgrading, the relative prospects for children of different origin may not improve. Since relative mobility closer reflects inequality of opportunity, it is central to our analysis. We find striking variation within countries in both upward and relative mobility, 2 and for both men and women. About two thirds of the variance in upward rank mobility, and half of the variance in relative rank mobility, occurs within countries rather than between them. Capital regions serve as hubs of upward mobility; sometimes, but not always, at the expense of relative mobility. In the highly mobile Nordic countries, there appears to be no such tradeoff, as capitals promote mobility of both kinds. Our analysis also contributes to the literature by explicitly incorporating women. While broad mobility patterns are similar across genders, mobility levels differ in most countries. Notably, in parts of southern and central Europe, women’s occupational attainment is shaped more strongly by social origin than that of men’s. Finally, we provide an exploratory analysis of potential mobility correlates, including human capital, labor market composition, demographics and other sociospatial characteristics. Upward mobility is greater in places with a high-skilled workforce, a large public sector, and bureaucratic organization. It is lower where unemployment is high, or a high proportion works in the primary sector, manufacturing, or services. Relative immobility is harder to predict, but there appears to be less mobility where occupations differ more widely in pay. We thereby add to the growing literature showing that in addition to childhood mechanisms, the local labor market faced as adult matters for mobility chances. 2 Previous literature 2.1 Intergenerational occupational mobility Occupational mobility has typically been the purview of sociologists (Breen and Jonsson, 2005), while economists have advanced the study of income mobility (Björklund and Jäntti, 2009). However, these boundaries are porous: to overcome measurement error, analyses of income mobility often impute income on occupations (Jácome et al., 2025; Kenedi and Sirugue, 2023), thereby effectively estimating occupational mobility. Moreover, recent work on income mobility has favored rankbased measures (Chetty et al., 2014; Engzell and Mood, 2023), which approach sociological notions of relative status (Bloome et al., 2018; Hout, 2018). Two traditions exist in the field of occupational mobility research. In the first, occupations are scaled in terms of gradational rank, providing a metric of occupational status (Ganzeboom et al., 1992). In the second approach, occupations are grouped into social classes. These classes are believed to reflect employment relations and might, in addition to current income level, provide a better representation of income stability, security and resources, than gradational measures (Bukodi and Goldthorpe, 2018; Goldthorpe, 1982). To ensure that our results are robust, we provide analyses speaking to both approaches. A fundamental distinction is that between absolute and relative mobility. Absolute mobility relates to the actual movements between origin and destination. It can be measured as the difference between parent and child occupational rank, or by whether your social class is different from the one you grew up in. Upward mobility from the bottom is often of particular interest, and we follow this focus in our study. Absolute mobility is to a large extent driven by variations in the occupational 3 structure, for example through changes in the labor market that create new occupations (often higher ranked) to replace declining industries (Breen, 2004; Erikson and Goldthorpe, 1992). In a regional context, absolute mobility can be achieved by moving between places with different occupational distributions; for example, from rural or industrial regions to more knowledge-intensive urban centers. Relative mobility (or social fluidity) refers to movements between origin and destination net of structural changes, capturing the dependency structure between parent and child. It can be measured using correlation coefficients for continuous measures or odds ratios for categorical measures. Relative mobility reflects social inequalities in the chances of reaching a given destination and is therefore considered an indicator of the relative openness of societies (Breen and Jonsson, 2005). Whereas absolute mobility accounts for most experienced mobility, relative mobility is of interest for the question of equality of opportunity. Even when upward mobility is high, relative immobility may dominate if opportunities are disproportionately concentrated among children from privileged backgrounds. We assess both in our analysis below. 2.2 Country differences Research on intergenerational income mobility finds higher mobility in Europe than in the US—southern Europe and the United Kingdom being the low-mobility exceptions—and the Nordic countries generally ranked as the most mobile (Blanden, 2013; Bratberg et al., 2017). This pattern is weaker for education and occupation, though the Nordics remain among the most mobile (Breen and Jonsson, 2005; Strömberg and Engzell, 2025). These trends are often attributed to comprehensive welfare states that equalize opportunities early in life (Corak, 2013; Erikson and Jonsson, 1996), although recent research also highlights labor market institutions (Landersø and Heckman, 2017; Mogstad et al., 2025). Analyses of income mobility tend to confirm the inverse relationship between inequality and mobility, both between and within countries (Havari et al., 2021; Narayan et al., 2018), whereas evidence for occupational mobility is more mixed (DiPrete, 2020). Inequality is commonly measured by the Gini coefficient, or occasionally top income shares. However, between-class inequality, indicating the resource distance between classes, has been argued to provide a better measure of inequality in relation to occupational mobility (Hertel and Groh-Samberg, 2019). Between-class inequality better reflects structural barriers to mobility, as it captures the economic distance between social groups, whereas the Gini coefficient does not distinguish whether inequality stems from differences within or between classes. We follow this argument in our analysis below, by measuring between-class inequality in income. The central motivation behind much occupational mobility research is the industrialization thesis (Lipset and Zetterberg, 1959; Treiman, 1970), which posits that the increasing complexity of modern production will place greater emphasis on formal credentials. As education expands and becomes more accessible, social origin should become less important in determining individual attainment. As a result, both absolute and relative mobility should increase with economic development. The industrialization thesis has been questioned and, in many studies, 4 refuted (e.g., Breen, 1997, 2004; Bukodi et al., 2020; Erikson and Goldthorpe, 1992; Knigge et al., 2014a) although some evidence also supports it (e.g., Aydemir and Yazici, 2019; Ganzeboom et al., 1989; Herrala, 2023; Knigge et al., 2014b; Yaish and Andersen, 2012). While evidence on the link between occupational mobility and macro-level economic indicators is mixed, one aspect of the industrialization thesis appears to hold: educational expansion is a key driver of both absolute and relative mobility (Breen and Müller, 2020). Because education mediates the relationship between social origins and destinations, broader access can help level the playing field. Notably, the link between social origins and destinations weakens at higher education levels, suggesting that high-skilled labor markets function more meritocratically (Breen and Jonsson, 2007; Hout, 1988; Karlson, 2019). 2.3 Regional variation Recent work on subnational differences in mobility—the “geography of opportunity”—has broadened the search for explanations beyond national-level factors. Most of this research studies income mobility, spurred by increased availability of administrative tax data (Connelly et al., 2016; Grusky et al., 2019; Song and Coleman, 2020). Influential work by Chetty et al. (2014) shows striking disparities in opportunity between places in the US. These efforts have since been extended to European countries, including Denmark (Eriksen and Munk, 2020), France (Kenedi and Sirugue, 2023), Italy (Acciari et al., 2022; Güell et al., 2018), the Netherlands (Beekers, 2024), Norway (Carneiro et al., 2019), Sweden (Brandén, 2019; Heidrich, 2017), and Switzerland (Chuard and Grassi, 2020). What then, are the local factors that promote or hinder mobility? In line with country-level studies, regional research highlights education as a key driver of mobility (Biasi, 2023; Corak, 2020; Eriksen and Munk, 2020). The availability and quality of schooling are crucial in childhood (Chetty et al., 2014; Stadelmann-Steffen, 2012), but education also shapes labor markets at both individual and collective levels. Higher average education can create social externalities, influencing the demand for skilled labor and fostering more dynamic economies (Cermeño, 2019; Liu, 2015). Additionally, high-skilled labor markets may promote mobility through knowledge spillovers and agglomeration effects, where workers gain skills through professional interactions, and industry clusters generate broader career prospects. Labor markets also matter in their own right: local unemployment rates are negatively associated with upward mobility (Acciari et al., 2022; Eriksen and Munk, 2020; Kenedi and Sirugue, 2023). Economic fluctuations in industries like manufacturing, mining, and petroleum are known to impact mobility (Berger and Engzell, 2022; Bütikofer et al., 2025; Deutscher and Mazumder, 2020). Sectoral composition likewise predicts mobility, with heavy industry and manufacturing showing correlations ranging from positive to negative, depending on the growth of those industries (Buscha et al., 2021; Heidrich, 2017). In Sweden and Canada, lower mobility is instead found in places with a stagnant primary industry (Corak, 2020; Heidrich, 2017). Historical research similarly highlights local industrial change as driving mobility, with the rapid transition from farming to manufacturing being a 5 watershed (Berger et al., 2023; Paterson et al., 2025). Another set of potential determinants is demographic. Social disadvantage often clusters in places where immigrants live, yet their children’s mobility often surpasses that of the general population (Borgen et al., 2025; Boustan et al., 2025; Hermansen, 2016). Regions with larger shares of immigrants (Corak, 2020; Eriksen and Munk, 2020) or indigenous populations (Connolly et al., 2019; Deutscher and Mazumder, 2020) are associated with lower levels of mobility, as are regions with a larger share of adults without a high school diploma (Corak, 2020). Moreover, the incidence of single parents negatively predicts mobility in several countries (Acciari et al., 2022; Connor et al., 2023; Corak, 2020; Eriksen and Munk, 2020; Peterson et al., 2023). High inequality can limit mobility by restricting access to quality education, professional networks, and well-paying jobs for those from disadvantaged backgrounds (Brandén, 2019; Corak, 2013). In the US, higher levels of residential segregation, both by income and race, along with greater income inequality, are associated with lower mobility levels (Chetty et al., 2014). Beyond economic disparities, social institutions also play a crucial role in shaping mobility outcomes. Social capital—the strength of networks and community ties—can facilitate mobility by providing support, job referrals, and mentorship (Chetty et al., 2022; Coleman, 1988). Likewise, social trust and broader civic capital, including institutional quality and community engagement, can foster environments where upward mobility is more attainable (Acciari et al., 2022; Peterson et al., 2023). Urbanization presents a mixed picture: while cities often offer greater economic opportunities and access to diverse networks, they can also exacerbate segregation and inequality, limiting mobility for lower-income groups (Breen and In, 2023; Connor and Storper, 2020; Friedman and Macmillan, 2017). Research from the US documents a rural advantage in upward mobility, where young men in particular benefit from stronger household and community supports, while women remain constrained by traditional gender norms (Connor et al., 2023). Finally, exposure to violent crime may hamper mobility by disrupting education, weakening social cohesion, and reducing economic investment, although evidence for this link is more consistent in the US than in Europe (Acciari et al., 2022; Kenedi and Sirugue, 2023; Manduca and Sampson, 2019; Sharkey and Torrats-Espinosa, 2017). 3 Theoretical expectations In this section, we outline our expectations regarding the influence of various measured predictors on mobility. There is a difference between absolute upward mobility, measuring the rate at which children from poorer families rise above their origins, and relative mobility, measured as rank correlations or odds ratios and capturing the overall similarity in status between parents and children. Given that most research on regional mobility disparities focuses on upward mobility, we formulate our hypotheses with this in mind while noting that these expectations may or may not extend to relative mobility. We expect that higher levels of human capital in a region will correlate with higher upward mobility. Education and training expand the kinds of work avail6 able, which can help individuals improve their social standing. Furthermore, access to better education resources, as indicated in our analysis by the number of teachers per capita, should be beneficial for upward mobility. Conversely, in places where most people have limited schooling, opportunities of this kind are likely to be scarcer. These patterns may reflect both individual-level advantages of an education and broader local economic conditions. Given the exploratory nature of our analysis, we do not attempt to disentangle these explanations. We expect that high unemployment rates will be negatively associated with upward mobility. Moreover, regions with a higher proportion of employment in low-skill sectors such as agriculture or manual labor (e.g., manufacturing) are likely to experience lower upward mobility, as these industries tend to offer fewer career opportunities. Conversely, regions with a higher proportion of employment in professional, technical, and service occupations should exhibit higher upward mobility. For upward mobility, these correlations are to some extent mechanical: more high-status jobs simply create more room at the top. It remains an open question whether these patterns extend to relative mobility, which depends on who actually gains access to those opportunities. We also measure the contribution of the public sector and bureaucratic roles to employment, which we expect to relate positively to mobility. Classically, Weber (1978) argued that bureaucracy could facilitate social mobility by providing clear rules and procedures for advancement. The relationship between immigration and mobility is ambiguous. Regions with larger immigrant populations may experience lower mobility overall, due to potential social and economic marginalization. This is what previous literature, focusing on childhood conditions, has tended to find (Corak, 2020; Eriksen and Munk, 2020). However, we measure location in adulthood, which complicates this expectation. Immigrants often seek upward mobility for their children and cluster in urban centers where opportunities are more abundant. Therefore, immigrant presence may correlate positively or negatively with mobility. We expect higher levels of single parenthood and larger household sizes to correlate with lower upward mobility. Single-parent households often face economic hardships that limit their children’s opportunities. Similarly, large families are associated with less investment per child (Gibbs et al., 2016), and typically also a society where women shoulder a larger part of the household burden (Menta and Lepinteur, 2021). We may expect these family factors to be especially important for women, although research also finds that young men are more vulnerable to disrupted family structures (Connor et al., 2023). Moreover, we investigate the extent of three-generation households and the median age of the population. Threegeneration households may foster stronger family bonds but could also increase occupational similarity between parents and children. Additionally, aging regions may indicate economic stagnation, further hindering mobility. Higher levels of inequality within a region, as measured by income differences between social classes (Hertel and Groh-Samberg, 2019), are expected to correlate negatively with mobility outcomes. Greater inequality not only raises the incentives for those at the top to maintain their position but also reinforces social barriers through differences in resources. In contrast, strong social capital and high levels of social trust are expected to be positively associated with upward mobility. So7 cial networks and trust have the potential to disperse opportunities more widely through job referrals, mentorship, and social support. Similarly, we expect areas with high civic engagement (e.g., political participation) to foster more inclusive environments, where upward mobility is more attainable. Urbanization may have a mixed effect on mobility. While cities often offer greater access to diverse job markets and educational opportunities, they can also amplify inequalities and social segregation, which may limit mobility for some groups. We expect urbanization to be associated with higher upward mobility, but not necessarily with greater relative mobility, due to the persistence of urban poverty and inequality. Additionally, higher levels of violent crime are expected to correspond with lower upward mobility, as crime can undermine social cohesion, weaken trust, and discourage investment in education and local communities. 4 Measuring intergenerational mobility We measure intergenerational mobility in terms of occupational rank and social class. For each, we distinguish between absolute upward and relative mobility, and between women and men. To measure rank mobility, we first create percentile scores in the national distribution of parent and child occupational status, described in greater detail below. We then estimate the following linear model: Rit =αj+βjRit−1+εit.(1) where Riis the respondent’s rank, Rit−1is the parental rank, αjand βjare parameters that vary at the level of local areas and εit is a residual error term. Both αjand βjare of interest as they capture, respectively, upward and relative mobility. In the following, we will call αjupward rank mobility and βjrank immobility. Upward rank mobility αjis the expected rank of respondents whose parents are at the bottom of the distribution. Higher values indicate that children born in the bottom can expect to attain a higher rank, and thus a higher value means more mobility. Rank immobility βj, also known as the rank-rank slope, captures the strength of the association between parent and child status. A slope of 0.20 indicates that parents on average pass on 20% of their (dis)advantage in occupational rank. The stronger βjis, the lower the intergenerational mobility, as it means that individuals can expect to end up on a rung close to where they were born. Within a country as a whole, or in the full sample, the relationship between the intercept and slope is deterministic, since the expected value at the median of parent rank must be approximately 50. This constraint does not hold at the subnational (regional) level, where there may be meaningful variation in the distribution of national occupational ranks. To obtain reliable estimates at a local level, we estimate Equation 1 using multilevel, or mixed, models (Snijders and Bosker, 2011). In principle, it is possible to obtain estimates of our parameters by running separate regressions for each region. The problem is that this will yield extreme estimates for some regions due to sampling variation (Heidrich, 2017). The multilevel solution starts from the assumption that the relationship of interest consists of a “fixed” component, common across all 8 (a) Upward rank mobility, men (b) Upward rank mobility, women Figure 2: Geographic variation in upward rank mobility (rank-rank intercept). Higher values indicate more mobility. Results for class mobility in Appendix Figure A8. providing an alternative representation of the intergenerational association. 6.1 The geography of intergenerational mobility Next, we turn to regional variation. Because the geographical patterns are similar for occupational rank and social class mobility, we concentrate on rank mobility, separating between upward mobility and relative immobility. Similar results for social class are shown in Figures A8–A9 of the Appendix. Figure 2 plots the variation in upward mobility for men and women, measured as the intercept (upward rank mobility). Higher values indicate more upward mobility. For men, the highest levels of upward rank mobility are found in some of the Nordic countries, with rates of around 46. For women, the Nordic countries are the most upwardly mobile, followed closely by parts of Ireland and some of the Baltic countries, with rates of around 43. The lowest rates for men are found in Ireland, around 28, and for women in southern Spain, around 20. These patterns follow known country differences in mobility (Bukodi et al., 2020). Yet, variation is mostly continuous, rather than clearly delineated by national borders. Upward rank mobility is systematically more common in capital regions, especially so in the Nordic countries, where Oslo, Stockholm, and Helsinki consistently rank among the ten most mobile regions in Europe. Berlin and Paris also stand out from their surrounding regions, and rank among the 20 most mobile regions. This relationship is likely explained in part by migration from provinces to capital regions, something we are unable to test with our data. In several other countries, the capital is visibly more mobile than the provinces, but does not enter among 15 (a) Relative rank immobility, men (b) Relative rank immobility, women Figure 3: Geographic variation in relative rank immobility (rank-rank slope). Higher values indicate less mobility. Results for class mobility in Appendix Figure A9. top ranking regions overall, by virtue of the general level of mobility in the country being lower. In Figure 3, we turn to relative rank immobility, measured as the slope of the rank-rank association. In contrast to upward rank mobility, higher values here indicate less mobility. Rank immobility ranges from a low of 0.23 for men in parts of the Baltic states and the Netherlands, with similar values for women in the Netherlands and Scandinavia. The highest rank immobility, around 0.45, is found in parts of Poland for men, and for women in Poland, Croatia, Hungary, Italy and southern Spain. The markedly higher relative immobility among women in much of southern and central Europe is the most striking gender difference in our data. How do these numbers compare for social class mobility? Figure A8 in the Appendix plots the variation in upward class mobility, measured as the proportion upwardly mobile from the working class. Higher values indicate more upward mobility. The pattern reflects the results shown in rank mobility, with higher levels of upward class mobility in the Netherlands and the Nordic countries (around 60% for both men and women). Additionally, the London region shows high levels of upward class mobility for men (59%), but not as high for women (50%). The lowest levels for men are found in parts of Bulgaria and Hungary (12%) and for women in parts of Bulgaria and Portugal (around 20%). Similar to rank mobility, upward class mobility is more common in capital regions, and especially so in the Nordic capitals. Figure A9 in the Appendix plots class immobility, measured as odds ratios, where higher values indicate more immobility. The lowest levels of class immobility are found mainly in the Nordic countries, as well as Estonia and the Netherlands, 16 with odds ratios around 2. The highest levels of immobility are found in Spain, Hungary and Poland (odds ratio around 6–7). The patterns of class immobility are similar to those of rank immobility, but with the exception of Northern Portugal being an outlier in class immobility for women, with odds ratios as high as 8. In mapping the landscape of relative rank and class mobility, capitals no longer stand out as hubs of mobility. In fact, in several examples such as Paris, Madrid, and Lisbon, inequality of opportunity is appreciably stronger in the capital than in the surrounding regions. Herein lies a potential dilemma for policy-makers: how can cities function as engines of opportunity without the attendant inequalities in who seizes those opportunities? Interestingly, this inverse relationship is not evident among Northern European capitals, which manage to achieve high upward mobility without sacrificing on relative mobility. How does this within-country variation compare to country differences that have been the focus of much previous research? Appendix Table A7 presents intraclass correlations reflecting how similar outcomes are for different regions within a country, that is, the proportion of variance captured by the between-country level. For rank mobility, this figure varies from about a third for upward rank mobility (men 0.29, women 0.38) to closer to half for relative rank immobility (men 0.37, women 0.54). This indicates that as much as half of variation in relative mobility, and two thirds of variation in absolute mobility, occurs between regions within countries. Class mobility varies more systematically between countries, but even here, between a third and two fifths of variation is within country. 6.2 Correlates of intergenerational mobility What explains local variation in intergenerational mobility? To answer this question, we estimate bivariate area-level correlations between each mobility metric and the comprehensive set of covariates described above. We estimate two sets of models, first using all geographic variation and then purging all between-country variation from explanatory and outcome variables, to capture within-country correlations only. Here, too, we concentrate on rank mobility, while showing similar results for social class in Figure A12 of the Appendix. We stress that these associations are ecological and suggestive, and we make no attempt at causal or even multivariate modeling. Our aim is to provide tentative evidence for future research to explore in greater depth. Figure 4, top panels, shows correlates of upward rank mobility (the rank-rank intercept) for men and women, respectively. Upward mobility is significantly correlated with most of the measures in our data and several human capital indicators show a substantial positive correlation with upward mobility. Most of these do not diminish in size or significance once we make use of the within-country variation only, thus confirming our hypothesis of a link between high levels of human capital and upward mobility. In places with a larger proportion of high-school dropouts, levels of upward mobility are lower. Labor market composition is another significant predictor that, following our expectations, shows lower levels of upward mobility in places with a large share of unemployed, and large primary and manufacturing sectors. Higher upward mobility 17 A. Human capital Average schooling Percent college Professionals STEM workers Teachers per capita Percent less than HS B. Labor market Unemployed Primary sector Manufacturing Services Public sector Bureaucracy C. Demographics Immigration Single mothers 3-gen households Household size Median age D. Socio-spatial Inequality Social capital Social trust Civic capital Urbanization Violent crime -1 -.5 0 .5 1 Correlation Total Within-country (a) Upward rank mobility, men A. Human capital Average schooling Percent college Professionals STEM workers Teachers per capita Percent less than HS B. Labor market Unemployed Primary sector Manufacturing Services Public sector Bureaucracy C. Demographics Immigration Single mothers 3-gen households Household size Median age D. Socio-spatial Inequality Social capital Social trust Civic capital Urbanization Violent crime -1 -.5 0 .5 1 Correlation Total Within-country (b) Upward rank mobility, women A. Human capital Average schooling Percent college Professionals STEM workers Teachers per capita Percent less than HS B. Labor market Unemployed Primary sector Manufacturing Services Public sector Bureaucracy C. Demographics Immigration Single mothers 3-gen households Household size Median age D. Socio-spatial Inequality Social capital Social trust Civic capital Urbanization Violent crime -.5 -.25 0 .25 .5 Correlation Total Within-country (c) Relative rank immobility, men A. Human capital Average schooling Percent college Professionals STEM workers Teachers per capita Percent less than HS B. Labor market Unemployed Primary sector Manufacturing Services Public sector Bureaucracy C. Demographics Immigration Single mothers 3-gen households Household size Median age D. Socio-spatial Inequality Social capital Social trust Civic capital Urbanization Violent crime -.5 -.25 0 .25 .5 Correlation Total Within-country (d) Relative rank immobility, women Figure 4: Correlates of intergenerational mobility. Correlation coefficient and 95% confidence interval, standard errors clustered at the country level. Results for class mobility in Appendix Figure A12. 18 is found in places with large public sectors and more bureaucratic organizations. For men, the correlation with public-sector employment is only significant between and not within countries, whereas for women it is significantly correlated both between and within countries although substantively small. The size of the service sector shows an inconsistent pattern: positively correlated with upward mobility for men, but only between countries, whereas it is negatively correlated within countries for women. Turning to demographic covariates, upward mobility is positively correlated with immigration and negatively correlated with median age; the latter only within countries, which might be an indication of stagnating regional development. Interestingly, the share of single mothers is positively correlated with women’s upward mobility both within and across countries, a pattern not observed for men. One possible interpretation is that regions supporting single mothers may also provide more inclusive institutions or services that facilitate women’s mobility, echoing earlier results that the dominance of two-parent households benefit the mobility of men over women (Connor et al., 2023). Next, we turn to indicators of inequality, urbanization, social capital, and trust. We have already noted that upward mobility is higher in capital regions, and this is evident in a positive correlation between urbanization and upward mobility, which aligns with our expectations. Somewhat surprisingly, violent crime is also positively correlated with upward mobility, most likely explained by crime being more prevalent in urban areas. Between-class income inequality is unrelated to upward mobility for either men or women. Lastly, and following our predictions, social capital, civic capital, and social trust are positively correlated with upward mobility, but mostly for men and between countries. We caution the reader that the amount of within-country variation in these variables is low (Appendix Table A7). In sum, upward mobility is associated with a range of human capital, labor market, demographic, and socio-spatial covariates. This picture changes once we turn to the rank-rank slope, measuring rank immobility or the strength of the association between parent and child occupation (Figure 4, bottom panels). Although some of the above covariates explain variation between countries, these associations turn non-significant when restricting variation to that within countries. One notable exception, however, is between-class income inequality which is negatively associated with relative mobility within countries for both men and women. Thus our results confirm previous research showing that the inverse relationship between inequality and (relative) mobility holds within and not just between countries (Chetty et al., 2014; Güell et al., 2018; Havari et al., 2021). While human capital variables show no clear association with relative immobility among men, they reveal a surprising pattern for women. As expected, across countries, higher educational attainment is linked to greater equality of opportunity. However, within countries, regions with higher education levels exhibit greater intergenerational persistence for women, a pattern that challenges the conventional view of education as an equalizer (Breen and Jonsson, 2007; Hout, 1988). It is more consistent, however, with observations from gender scholarship suggesting that women’s entry into traditionally male-dominated domains like higher education and professional occupations has been uneven, disproportionately benefiting 19 those from privileged backgrounds (England, 2010). In contrast, men’s relative mobility appears less influenced by regional educational composition, suggesting that their intergenerational reproduction in the labor market depends less on education. 6.3 Comparison with previous work How do our results compare with previous work on regional mobility variation within countries? A related study is Betthäuser et al. (2021), which examines educational mobility using the same dataset and regions. They estimate intergenerational risk ratios for tertiary education, most comparable to the odds ratios for class immobility in our analysis. Some within-country patterns align across studies; for example, lower mobility in western Germany and northern Portugal appears in both education and occupation. In contrast, countries like France and Poland show divergent or even reversed patterns. There are also notable cross-country differences. Britain and Germany, for instance, are high-mobility cases in Betthäuser et al. (2021), but appear more middling or low-mobility in our results. A likely explanation is that tertiary attainment captures limited variation, overlooking differences in field of study, institutional prestige, or program duration, all of which matter for occupational outcomes (Gerber and Cheung, 2008; Reimer and Thomsen, 2019). Several studies have also estimated regional variation in income mobility within one country (e.g., Acciari et al., 2022; Brandén, 2019; Bütikofer et al., 2025; Güell et al., 2018; Heidrich, 2017; Kenedi and Sirugue, 2023) or, in some cases, occupational mobility (Bell et al., 2023; Breen and In, 2024a,b; Buscha et al., 2021). These studies typically examine variation by region of origin rather than residence, and typically aggregate results for men and women or report only on men. Our findings for capital regions receive support in some, but not all, earlier work. Previous studies find that capital regions can be sites of high absolute (Bell et al., 2023; Kenedi and Sirugue, 2023) and low relative mobility (Brandén, 2019; Heidrich, 2017), but the patterns by country are not always the same as ours. Some of these differences may be driven by the distinction between occupation and income. Capital regions tend to be economically dynamic, offering diverse labor markets and opportunities that can help individuals move beyond their geographic or class origins. At the same time, they often have a highly stratified upper tier, which may be better captured by detailed income data than by occupational classifications. More research is needed on under what conditions, and for whom, capitals serve as engines of mobility (Breen and In, 2023; Friedman and Macmillan, 2017). In several cases, we find that earlier results in the literature replicate mainly with respect to men and upward mobility. This applies to previously known north–south gradients in England (high mobility in the south, Bell et al., 2023) and Italy (low mobility in the south, Acciari et al., 2022; Güell et al., 2018), but also to the lower mobility in eastern Germany (Dodin et al., 2024) or higher mobility in Norwegian oil regions (Bütikofer et al., 2025). In each of these cases, we reproduce patterns in upward mobility for men, but not in relative mobility or for women. Future work on regional mobility differences would therefore be well-advised to pay closer attention to relative mobility and gender differences. Our finding of higher relative mobility in east Germany is, however, consistent with previous research using social class 20 to predict children’s education or occupation (Betthäuser, 2019; Pollak and Müller, 2002), showing that occupational and income mobility are different concepts that need not follow the same patterns (Breen et al., 2016). 7 Discussion and conclusion In this paper, we have examined local variation in intergenerational mobility by using the European Social Survey. We show that intergenerational mobility varies at least as much within countries as between them. Capital regions are hubs of upward mobility, but not generally of relative mobility. While upward mobility is related to a range of characteristics of local labor markets, including industrial composition and human capital, relative mobility turns out to be more idiosyncratic and harder to predict. However, relative mobility is related to class differences in income, such that high levels of between-class segmentation make it harder to climb the social ladder. Patterns also differ notably for women and men, with betweencountry differences generally being more marked for women’s opportunities. Much previous research has focused on the mechanisms of mobility in childhood. This paper adds to mounting evidence showing that adult residence and labor market conditions also play a role. Our findings, along with this wider literature, suggest important implications for social policy. To improve mobility, it is essential to focus not only on ensuring equal opportunities for children but also on strengthening labor markets across Europe. However, further research is needed to disentangle the role of childhood and adult residence, especially taking self-selection into migration into account. This may also help explain why mobility outcomes vary more within countries for men, as couples are more likely to relocate for the man’s career (Brandén and Haandrikman, 2019; Jayachandran et al., 2024), making selective migration a potentially more influential factor here. At the same time, women’s mobility shows large differences between countries, particularly in relative immobility. While we cannot conclusively determine the reasons for this, it may be related to differences in work and gender norms, as well as the availability of institutional support for women’s careers. In particular, relative immobility for women is extreme in parts of southern and central Europe. With more traditional gender roles, women face stronger social expectations to stay close to home, take on caregiving responsibilities, or prioritize family over career. In such contexts, women may also encounter more segmented labor markets with limited access to jobs, particularly in male-dominated spheres. This explains why women’s attainment would be lower in general, but not the strong dependence on occupational origins. Further research should aim to better understand these gendered mobility patterns and their underlying factors. To understand local variation in intergenerational mobility, we explore how arealevel mobility metrics correlate with a range of regional characteristics. By examining both betweenand within-country variation, we find that upward mobility is linked to indicators of human capital, such as low dropout rates and high educational attainment. These patterns persist even when focusing on within-country variation. Labor market composition also plays a role: areas with more bureaucratic roles and 21 (for women) large public sectors tend to have higher upward mobility, while regions dominated by primary and manufacturing sectors show lower mobility. Additionally, demographic factors such as younger populations and higher levels of immigration are associated with more upward mobility. We have not attempted to disentangle whether these associations reflect the characteristics of individuals who are more or less mobile, or the characteristics of places themselves. For instance, higher education may benefit individual mobility but could also influence local labor markets through increased productivity and demand for skilled labor. When shifting focus to relative immobility, the associations weaken within countries, except for a clear negative relationship with between-class income inequality. This supports the idea, well documented in prior research, that greater inequality tends to reduce mobility, even at the regional level. Comparisons with earlier studies show that some patterns replicate, especially for men and upward mobility, such as north–south divides in the UK and Italy or higher mobility in Norway’s oil-rich areas. However, patterns are different for relative mobility and the mobility of women. Deviations from previous studies may also result from the varying focus on occupation and income, illustrating the need to study these two dimensions not only separately but also in combination. Future research should also aim to better understand the interplay between national institutions and regional labor markets. While national factors may not directly explain regional differences in a statistical sense, this does not mean that national policies do not influence regional outcomes. Just as decentralized school policy can lead to large variations in educational quality and equality across regions (Biasi, 2023; Stadelmann-Steffen, 2012), national industrial policies or labor market interventions can create significant regional differences in opportunities. These kinds of national policies may not lend themselves to statistical modeling of withincountry variation, but they can still have an impact on regional differences. Our contribution bridges recent research based on administrative data that explores subnational variation in detail, with earlier work comparing countries using large-scale representative survey data. Administrative data are collected by national statistical agencies with varying regulations, protocols, coverage, time spans, and quality. As a result, recent research on regional variation often uses “islands of data,” which capitalize on the specifics of a given dataset to provide a detailed view of one country, but without the comparative element. Survey research, on the other hand, has until recently lacked sufficient sample sizes to examine detailed subnational variation, but the maturation of long-running surveys like the European Social Survey has made this increasingly possible. Our work offers a template for how researchers can combine the strengths of each approach. To facilitate further work, we make our code and estimates available to researchers wishing to explore intergenerational mobility and its correlates. Acknowledgements Funding was generously provided by the European Research Council grant no. 101165962 (MaMo). We thank Bastian Betthäuser, Caspar Kaiser, and Nhat An Trinh for sharing R code that aided our analysis. Earlier versions of this work were presented at the 2024 Workshop on Spatial Inequalities in Europe at University College London and the 2024 European Consortium for So22 ciological Research (ECSR) conference in Barcelona. We thank the audiences on these occasions, as well as the editorial board and three anonymous reviewers for European Societies whose comments greatly improved the manuscript. AI use disclosure AI-based tools were used for some copyediting tasks. All analyses and data preparation were performed without the use of AI. Author contributions Both authors contributed equally to this work. Data availability Source data are available for download from the European Social Survey. The authors’ dataset containing aggregate regional level statistics, as well as code to reproduce all tables and figures, are available at https://osf. io/24w36/. Declaration of interest statement The authors declare no conflicts of interest. Ethical approval This study uses secondary data from the European Social Survey, a publicly available dataset collected under strict ethical guidelines and informed consent protocols established by the ESS. Only anonymized data were accessed, and no personal identifiers were used in any part of the analysis. No additional ethical approval was needed. Supplements Supplemental data for this article can be accessed online. References Acciari, P., Polo, A., and Violante, G. L. (2022). And yet it moves: Intergenerational mobility in Italy. American Economic Journal: Applied Economics, 14(3):118–63. Ahsan, M. N., Emran, M. S., Jiang, H., Murphy, O., and Shilpi, F. (2022). When measures conflict: Towards a better understanding of intergenerational educational mobility. Available at SSRN. Aydemir, A. B. and Yazici, H. (2019). Intergenerational education mobility and the level of development. European Economic Review, 116:160–85. Beekers, L. (2024). Sources of regional variation in intergenerational mobility: Evidence from the Netherlands. CentER Discussion Paper 2024-015, CentER, Center for Economic Research. Bell, B., Blundell, J., and Machin, S. (2023). Where is the land of hope and glory? The geography of intergenerational mobility in England and Wales. Scandinavian Journal of Economics, 125(1):73–106. Beller, E. (2009). Bringing intergenerational social mobility research into the twenty-first century: Why mothers matter. American Sociological Review, 74(4):507–528. 23 Berger, T. and Engzell, P. (2022). Industrial automation and intergenerational income mobility in the United States. Social Science Research, 102686. Berger, T., Engzell, P., Eriksson, B., and Molinder, J. (2023). Social mobility in Sweden before the welfare state. Journal of Economic History, 83(2):431–463. Betthäuser, B. A. (2019). The effect of the post-socialist transition on inequality of educational opportunity: Evidence from German unification. European Sociological Review, 35(4):461–473. Betthäuser, B. A., Kaiser, C., and Trinh, N. A. (2021). Regional variation in inequality of educational opportunity across Europe. Socius, 7(23780231211019890). Biasi, B. (2023). School finance equalization increases intergenerational mobility. Journal of Labor Economics, 41(1):1–38. Björklund, A. and Jäntti, M. (2009). Intergenerational income mobility and the role of family background. In Salverda, W., Nolan, B., and Smeeding, T., editors, Oxford Handbook of Economic Inequality, pages 491–521. Oxford University Press. Blanden, J. (2013). Cross-country rankings in intergenerational mobility: A comparison of approaches from economics and sociology. Journal of Economic Surveys, 27(1):38–73. Bloome, D., Dyer, S., and Zhou, X. (2018). Educational inequality, educational expansion, and intergenerational income persistence in the United States. American Sociological Review, 83(6):1215–1253. Borgen, S. T., Borgen, N. T., and Zachrisson, H. D. (2025). Ethnic minorities in disadvantaged neighborhoods: Differential effects on educational outcomes. Social Science Research, 127:103152. Boustan, L., Jensen, M. F., Abramitzky, R., Jácome, E., Manning, A., Pérez, S., Watley, A., Adermon, A., Arellano-Bover, J., Åslund, O., et al. (2025). Intergenerational mobility of immigrants in 15 destination countries. Technical report, National Bureau of Economic Research. Brandén, M. and Haandrikman, K. (2019). Who moves to whom? Gender differences in the distance moved to a shared residence. European Journal of Population, 35:435–458. Brandén, G. (2019). Does inequality reduce mobility? The Great Gatsby curve and its mechanisms. Working Paper, Institute for Evaluation of Labour Market and Education Policy (IFAU), Uppsala. Bratberg, E., Davis, J., Mazumder, B., Nybom, M., Schnitzlein, D. D., and Vaage, K. (2017). A comparison of intergenerational mobility curves in Germany, Norway, Sweden, and the US. Scandinavian Journal of Economics, 119(1):72–101. 24 Strömberg, E. and Engzell, P. (2025). How robust are country rankings in educational mobility? Sociological Science, forthcoming. Stuhler, J. (2018). A review of intergenerational mobility and its drivers. Publications Office of the European Union, Luxembourg. Thaning, M. and Hällsten, M. (2020). The end of dominance? Evaluating measures of socio-economic background in stratification research. European Sociological Review, 36(4):533–47. Treiman, D. J. (1970). Industrialization and social stratification. Sociological Inquiry, 40:207–34. Vigna, N. (2023). Subjective social status in places that don’t matter: Geographical inequalities in France and Germany. European Societies, 25(5):693–720. Weber, M. (1978). Economy and Society: An Outline of Interpretive Sociology. University of California Press, Berkeley, CA. Originally published in German in 1922. Winship, C. and Radbill, L. (1994). Sampling weights and regression analysis. Sociological Methods & Research, 23(2):230–257. Yaish, M. and Andersen, R. (2012). Social mobility in 20 modern societies: The role of economic and political context. Social Science Research, 41(3):527–38. Zwysen, W. (2016). Crowding out of disadvantaged young adults in Germany: Background matters depending on local labour market. European Sociological Review, 32(5):662–674. 31 Appendix for: The Geography of Intergenerational Mobility in Europe (for online publication only) 1 Table A1: Availability of ESS rounds by country. Country R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 Albania ✓ Austria ✓ ✓ ✓ ✓ ✓ ✓ Belgium ✓ ✓ ✓ ✓ ✓ ✓ Bulgaria ✓ ✓ ✓ ✓ ✓ ✓ Croatia ✓ ✓ ✓ ✓ Cyprus ✓ ✓ ✓ ✓ ✓ Czech Republic ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ Denmark ✓ ✓ ✓ ✓ ✓ ✓ ✓ Estonia ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ Finland ✓ ✓ ✓ ✓ ✓ ✓ ✓ France ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ Germany ✓✓✓✓✓✓✓✓✓ Greece ✓ ✓ ✓ ✓ Hungary ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ Iceland ✓ ✓ ✓ ✓ ✓ Ireland ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ Italy ✓ ✓ ✓ ✓ Kosovo ✓ Latvia ✓ ✓ Lithuania ✓ ✓ ✓ ✓ ✓ ✓ Luxembourg ✓ ✓ Montenegro ✓ ✓ Netherlands ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ Norway ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ Poland ✓✓✓✓✓✓✓✓✓ Portugal ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ Romania ✓ Serbia ✓ Slovak Republic ✓ ✓ ✓ ✓ ✓ ✓ ✓ Slovenia ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ Spain ✓ ✓ ✓ ✓ ✓ ✓ ✓ Sweden ✓✓✓✓✓✓✓✓✓ Switzerland ✓ ✓ ✓ ✓ ✓ ✓ ✓ Ukraine ✓ ✓ ✓ ✓ ✓ United Kingdom ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ 2 Table A2: Percent service class, observations and regions per country. Percent service Country Parents Child Observations Regions Albania 20 37 585 1 Austria 35 47 10310 9 Belgium 35 55 7687 11 Bulgaria 22 34 10622 28 Croatia 26 40 3970 6 Cyprus 18 43 3491 1 Czech Republic 32 43 13890 14 Denmark 43 52 9311 5 Estonia 35 41 13919 7 Finland 27 48 11699 8 France 35 53 12519 19 Germany 34 54 20051 16 Greece 20 37 5383 13 Hungary 18 35 12500 9 Iceland 45 57 3442 1 Ireland 26 46 12621 8 Italy 23 46 5289 5 Kosovo 22 45 420 1 Latvia 25 37 2426 6 Lithuania 18 37 8605 17 Luxembourg 29 48 1999 1 Montenegro 24 42 1046 1 Netherlands 39 59 14645 12 Norway 45 54 13796 10 Poland 22 39 11266 16 Portugal 17 32 10950 5 Romania 15 36 1124 8 Serbia 25 40 1211 1 Slovak Republic 25 40 8655 8 Slovenia 29 48 8103 16 Spain 21 38 8427 7 Sweden 42 52 13506 8 Switzerland 36 60 9133 7 Ukraine 28 40 7695 26 United Kingdom 39 50 16172 12 Total 31 46 296468 323 3 Table A3: Coding of respondent social class. ESeC 2002, 2004, 2006, 2008, 2010 ESeC 2012, 2014, 2016, 2018 This study •Large employers, higher managers / professional •Lower managers / professional, higher supervisors •Intermediate occupations •Large employers, higher managers / professional •Lower managers / professional, higher supervisors •Intermediate occupations Service class •Lower supervisors and technicians •Lower sales and service •Lower technical •Routine •Small employers and self-employed (agriculture incl.) •Lower supervisors and technicians •Lower sales and service •Lower technical •Routine Working class 4 Table A4: Coding of parent social class. Parent occupation 2002, 2004, 2006 Parent occupation 2008, 2010, 2012, 2014, 2016, 2018 This study •Traditional professional •Modern professional •Clerical and intermediate •Senior manager or administrators •Middle or junior managers •Professional and technical •Higher administrator •Clerical Service class •Technical and craft •Farmer •Semi-routine/manual/ service •Routine manual and service •Skilled worker •Semi-skilled worker •Unskilled •Service •Sales •Farmer Working class 5 Table A5: Coding of father’s occupational rank. 2002 2004-2006 2008-2018 Ireland, 2002 This study Semiroutine/manual/ service occupations Semiroutine/manual/ service occupations Farm worker Farmer 1 Routine manual and service occupations Routine manual and service occupations Unskilled worker Routine manual and service occupations 2 Technical and craft occupations Technical and craft occupations, farmer Semi-skilled worker Semiroutine/manual/ service occupation 3 Middle or junior managers Middle or junior managers Skilled worker Technical and craft occupations 4 Clerical and intermediate occupations Clerical and intermediate occupations Service occupations Traditional professional occupations 5 Senior manager or administrators Traditional professional occupations Sales occupations Clerical and intermediate occupations 6 Modern professional occupations Modern professional occupations Clerical occupations Modern professional occupations 7 Traditional professional occupations Senior manager or administrators Higher administrator occupations Middle or junior managers 8 — — Professional and technical occupations Senior manager or administrators 9 6 Table A6: Coding of mother’s occupational rank. 2002 2004-2006 2008-2018 Ireland, 2002 This study Semiroutine/manual/ service occupation Technical and craft occupations, farmer Farm worker Farmer 1 Routine manual and service occupations Semiroutine/manual/ service occupations Unskilled worker Technical and craft occupations 2 Technical and craft occupations Routine manual and service occupations Semi-skilled worker Routine manual and service occupations 3 Middle or junior managers Middle or junior managers Skilled worker Clerical and intermediate occupations 4 Senior manager or administrators Clerical and intermediate occupations Service occupations Semiroutine/manual/ service occupation 5 Clerical and intermediate occupations Senior manager or administrators Sales occupations Senior manager or administrators 6 Modern professional occupations Traditional professional occupations Clerical occupations Modern professional occupations 7 Traditional professional occupations Modern professional occupations Higher administrator occupations Traditional professional occupations 8 — — Professional and technical occupations Middle or junior managers 9 7 Table A7: Intra-class correlation coefficients at the country level for intergenerational mobility and covariates. (a) Unweighted Variable ICC Rank-rank intercept, men 0.291 Rank-rank intercept, women 0.378 Rank-rank slope, men 0.376 Rank-rank slope, women 0.541 Upward class mobility, men 0.665 Upward class mobility, women 0.538 Class immobility (odds), men 0.350 Class immobility (odds), women 0.541 Average schooling 0.774 Percent college 0.691 Professionals 0.493 STEM workers 0.542 Teachers per capita 0.580 Percent less than HS 0.913 Unemployed 0.785 Primary sector 0.550 Manufacturing 0.538 Services 0.602 Public sector 0.657 Bureaucracy 0.702 Immigration 0.549 Single mothers 0.424 3-gen households 0.816 Household size 0.913 Median age 0.480 Inequality 0.505 Social capital 0.882 Social trust 0.894 Civic capital 0.775 Urbanization 0.000 Violent crime 0.601 (b) Weighted Variable ICC Rank-rank intercept, men 0.288 Rank-rank intercept, women 0.304 Rank-rank slope, men 0.264 Rank-rank slope, women 0.522 Upward class mobility, men 0.624 Upward class mobility, women 0.534 Class immobility (odds), men 0.298 Class immobility (odds), women 0.488 Average schooling 0.809 Percent college 0.735 Professionals 0.523 STEM workers 0.516 Teachers per capita 0.519 Percent less than HS 0.906 Unemployed 0.504 Primary sector 0.562 Manufacturing 0.433 Services 0.579 Public sector 0.565 Bureaucracy 0.657 Immigration 0.387 Single mothers 0.333 3-gen households 0.709 Household size 0.577 Median age 0.421 Inequality 0.518 Social capital 0.894 Social trust 0.889 Civic capital 0.733 Urbanization 0.012 Violent crime 0.560 8 Table A8: Correlation matrix between intergenerational mobility and covariates (unweighted). Variables (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) (13) (14) (15) (16) (17) (18) (19) (20) (21) (22) (23) (24) (25) (26) (27) (28) (29) (30) (31) Rank-rank intercept, men 1.00 Rank-rank intercept, women 0.23 1.00 Rank-rank slope, men -0.03 -0.18 1.00 Rank-rank slope, women -0.11 -0.38 0.33 1.00 Upward class mobility, men 0.66 0.24 0.19 -0.32 1.00 Upward class mobility, women 0.38 0.68 0.04 -0.31 0.68 1.00 Class immobility (odds), men -0.15 -0.30 0.61 0.46 -0.25 -0.27 1.00 Class immobility (odds), women -0.09 -0.38 0.37 0.68 -0.18 -0.33 0.55 1.00 Average schooling 0.38 0.56 0.08 -0.26 0.49 0.63 -0.24 -0.24 1.00 Percent college 0.33 0.47 -0.05 -0.13 0.30 0.46 -0.22 -0.25 0.48 1.00 Professionals 0.58 0.54 0.15 -0.22 0.71 0.78 -0.13 -0.20 0.69 0.56 1.00 STEM workers 0.43 0.25 0.15 -0.13 0.52 0.50 -0.07 -0.08 0.43 0.26 0.57 1.00 Teachers per capita 0.32 0.28 0.03 -0.29 0.50 0.51 -0.18 -0.25 0.44 0.37 0.69 0.28 1.00 Percent less than HS -0.04 -0.25 -0.09 -0.09 0.02 -0.22 0.03 0.03 -0.45 -0.14 -0.19 -0.29 -0.02 1.00 Unemployed -0.25 -0.23 -0.14 0.21 -0.45 -0.43 0.04 0.09 -0.39 -0.16 -0.41 -0.41 -0.37 0.11 1.00 Primary sector -0.41 -0.33 -0.10 0.06 -0.37 -0.40 0.01 -0.02 -0.49 -0.19 -0.48 -0.46 -0.28 0.36 0.36 1.00 Manufacturing -0.34 -0.19 -0.09 0.20 -0.56 -0.44 0.11 0.13 -0.16 -0.29 -0.41 0.03 -0.32 -0.31 0.06 -0.06 1.00 Services 0.19 -0.05 -0.07 -0.41 0.17 -0.07 -0.14 -0.26 0.13 0.04 0.15 0.02 0.31 0.30 -0.13 -0.10 -0.27 1.00 Public sector 0.36 0.27 0.02 -0.39 0.58 0.56 -0.25 -0.29 0.53 0.37 0.73 0.35 0.84 -0.13 -0.28 -0.34 -0.31 0.35 1.00 Bureaucracy 0.46 0.23 0.24 -0.24 0.74 0.57 -0.08 -0.06 0.51 0.11 0.57 0.48 0.41 -0.04 -0.48 -0.47 -0.36 0.16 0.44 1.00 Immigration 0.26 0.27 0.09 -0.23 0.46 0.41 -0.10 -0.13 0.26 0.35 0.46 0.30 0.27 0.07 -0.21 -0.32 -0.36 0.12 0.26 0.46 1.00 Single mothers -0.12 0.14 -0.13 0.12 -0.34 -0.16 -0.04 0.03 0.09 0.30 -0.10 -0.05 -0.10 -0.17 0.11 -0.06 0.27 -0.09 -0.13 -0.22 0.09 1.00 3-gen households -0.30 -0.18 -0.13 0.33 -0.55 -0.48 0.10 0.11 -0.48 -0.02 -0.50 -0.39 -0.46 -0.12 0.59 0.38 0.26 -0.36 -0.47 -0.60 -0.33 0.19 1.00 Household size -0.17 -0.11 0.03 0.25 -0.21 -0.24 0.08 0.15 -0.33 -0.07 -0.35 -0.32 -0.33 -0.01 0.47 0.21 0.02 -0.35 -0.37 -0.37 -0.20 -0.04 0.69 1.00 Median age -0.11 -0.17 -0.24 -0.08 -0.28 -0.32 -0.13 -0.09 -0.26 -0.06 -0.21 -0.17 -0.08 0.15 0.28 0.24 0.27 0.11 0.08 -0.27 -0.18 0.20 0.16 -0.25 1.00 Inequality 0.13 -0.05 0.20 0.09 0.22 0.10 0.16 0.18 0.02 -0.15 0.13 0.10 -0.01 0.34 0.04 -0.10 -0.22 0.10 0.00 0.27 0.09 -0.15 -0.25 -0.07 -0.16 1.00 Social capital 0.36 -0.02 0.02 -0.34 0.46 0.17 -0.12 -0.14 0.09 0.07 0.34 0.29 0.34 0.18 -0.02 -0.27 -0.30 0.49 0.48 0.41 0.25 -0.24 -0.27 -0.10 0.05 0.19 1.00 Social trust 0.36 0.29 0.11 -0.46 0.60 0.57 -0.21 -0.30 0.57 0.22 0.64 0.45 0.59 -0.08 -0.54 -0.34 -0.26 0.42 0.65 0.53 0.34 -0.21 -0.73 -0.49 -0.18 0.10 0.41 1.00 Civic capital 0.26 0.20 0.03 -0.42 0.56 0.43 -0.21 -0.26 0.34 0.10 0.44 0.25 0.48 0.10 -0.38 -0.17 -0.35 0.33 0.44 0.48 0.29 -0.33 -0.49 -0.21 -0.29 0.11 0.38 0.65 1.00 Urbanization 0.37 0.28 0.06 0.08 0.13 0.20 0.09 0.03 0.27 0.28 0.33 0.20 -0.02 -0.07 -0.03 -0.25 -0.28 0.05 -0.03 0.18 0.24 0.15 -0.05 -0.12 -0.22 0.08 0.01 -0.03 -0.17 1.00 Violent crime 0.36 0.21 -0.09 -0.27 0.38 0.33 -0.17 -0.22 0.31 0.28 0.43 0.35 0.32 0.25 -0.17 -0.16 -0.28 0.35 0.32 0.34 0.36 -0.01 -0.41 -0.29 -0.16 0.23 0.37 0.34 0.31 0.29 1.00 9 (a) Upward rank mobility, father-son (b) Upward rank mobility, mother-daughter Figure A10: Geographic variation in gender-specific upward rank mobility (rankrank intercept). Higher values indicate more mobility. (a) Relative rank immobility, father-son (b) Relative rank immobility, mother-daughter Figure A11: Geographic variation in gender-specific relative rank immobility (rankrank slope). Higher values indicate less mobility. 16 A. Human capital Average schooling Percent college Professionals STEM workers Teachers per capita Percent less than HS B. Labor market Unemployed Primary sector Manufacturing Services Public sector Bureaucracy C. Demographics Immigration Single mothers 3-gen households Household size Median age D. Socio-spatial Inequality Social capital Social trust Civic capital Urbanization Violent crime -1 -.5 0 .5 1 Correlation Total Within-country (a) Upward class mobility, men A. Human capital Average schooling Percent college Professionals STEM workers Teachers per capita Percent less than HS B. Labor market Unemployed Primary sector Manufacturing Services Public sector Bureaucracy C. Demographics Immigration Single mothers 3-gen households Household size Median age D. Socio-spatial Inequality Social capital Social trust Civic capital Urbanization Violent crime -1 -.5 0 .5 1 Correlation Total Within-country (b) Upward class mobility, women A. Human capital Average schooling Percent college Professionals STEM workers Teachers per capita Percent less than HS B. Labor market Unemployed Primary sector Manufacturing Services Public sector Bureaucracy C. Demographics Immigration Single mothers 3-gen households Household size Median age D. Socio-spatial Inequality Social capital Social trust Civic capital Urbanization Violent crime -.5 -.25 0 .25 .5 Correlation Total Within-country (c) Relative class immobility, men A. Human capital Average schooling Percent college Professionals STEM workers Teachers per capita Percent less than HS B. Labor market Unemployed Primary sector Manufacturing Services Public sector Bureaucracy C. Demographics Immigration Single mothers 3-gen households Household size Median age D. Socio-spatial Inequality Social capital Social trust Civic capital Urbanization Violent crime -.5 -.25 0 .25 .5 Correlation Total Within-country (d) Relative class immobility, women Figure A12: Correlates of intergenerational mobility, social class. Correlation coefficient and 95% confidence interval, standard errors clustered at the country level. 17 A. Human capital Average schooling Percent college Professionals STEM workers Teachers per capita Percent less than HS B. Labor market Unemployed Primary sector Manufacturing Services Public sector Bureaucracy C. Demographics Immigration Single mothers 3-gen households Household size Median age D. Socio-spatial Inequality Social capital Social trust Civic capital Urbanization Violent crime -1 -.5 0 .5 1 Correlation Total Within-country (a) Upward rank mobility, men A. Human capital Average schooling Percent college Professionals STEM workers Teachers per capita Percent less than HS B. Labor market Unemployed Primary sector Manufacturing Services Public sector Bureaucracy C. Demographics Immigration Single mothers 3-gen households Household size Median age D. Socio-spatial Inequality Social capital Social trust Civic capital Urbanization Violent crime -1 -.5 0 .5 1 Correlation Total Within-country (b) Upward rank mobility, women A. Human capital Average schooling Percent college Professionals STEM workers Teachers per capita Percent less than HS B. Labor market Unemployed Primary sector Manufacturing Services Public sector Bureaucracy C. Demographics Immigration Single mothers 3-gen households Household size Median age D. Socio-spatial Inequality Social capital Social trust Civic capital Urbanization Violent crime -.5 -.25 0 .25 .5 Correlation Total Within-country (c) Rank immobility, men A. Human capital Average schooling Percent college Professionals STEM workers Teachers per capita Percent less than HS B. Labor market Unemployed Primary sector Manufacturing Services Public sector Bureaucracy C. Demographics Immigration Single mothers 3-gen households Household size Median age D. Socio-spatial Inequality Social capital Social trust Civic capital Urbanization Violent crime -.5 -.25 0 .25 .5 Correlation Total Within-country (d) Rank immobility, women Figure A13: Correlates of intergenerational mobility (unweighted). Correlation coefficient and 95% confidence interval, standard errors clustered at the country level. 18