Employment trajectories in heterogeneous regions: Evidence from Germany

Advances in Life Course Research 40 (2019) 43–84

Contents lists available at ScienceDirect

Advances in Life Course Research journal homepage: www.elsevier.com/locate/alcr

Employment trajectories in heterogeneous regions: Evidence from Germany Matthias Dütsch a b

a,b,⁎

a

, Franziska Ganesch , Olaf Struck

a

T

Chair of Labour Studies, Otto-Friedrich University of Bamberg, Feldkirchenstraße 21, 96052 Bamberg, Germany Federal Institute for Occupational Safety and Health, Nöldnerstraße 40-42, 10317 Berlin, Germany

ARTICLE INFO

ABSTRACT

Keywords: Regional disparities Employment career Life course Germany Event history analysis

To what extent do regional characteristics influence employment trajectories? Do regional factors diversely affect the employment careers of different sociodemographic groups? By investigating these questions, we extend current life course research in two ways: First, from a conceptual perspective, we use approaches from regional economics in addition to established sociological labour market theories to gain insights into the effects of regional determinants on individual labour market outcomes. Second, from a methodological point of view, we conduct event history analyses based on a German dataset that contains information on individuals, firms and regions. Our results show that there are considerable regional heterogeneities regarding population density and the amount of human capital endowment, both of which influence working careers differently. Regional agglomeration predominantly offers opportunities in terms of employment trajectories, while regional human capital accumulation increases employment risks. Additionally, our findings indicate that group-specific inequalities with respect to employment careers can be weakened or even strengthened by regional frame conditions. Female and foreign employees benefit most from denser regions and from a higher human capital endowment. By contrast, the unemployment risks of workers who previously experienced unemployment periods during their working lives are increased by both of these regional characteristics. Findings regarding education level are mixed: Workers with occupational qualifications profit from regional agglomeration to a greater extent than do low or even generally qualified workers. However, a high local human capital endowment leads to skill segregation between vocationally trained and highly qualified employees.

1. Introduction The structure and functioning of labour markets play an important role in modern societies; both have a significant impact on how economic wealth is shared among members of a society, as income and derived contributions to the social security system are both based on paid work. Labour markets thus affect the structure of social inequalities by determining the employment opportunities and risks within the workforce as well as the employment options of the unemployed (Heinz, Huinink, & Weymann, 2009). For workers, stable employment is highly relevant because it provides both earnings and the opportunity to maintain and increase their human capital (Blossfeld, Mills, & Bernardi, 2006). Empirical findings paint a heterogeneous picture of employment stability in Germany. Research that examines developments up to the mid-1990s (Erlinghagen, 2004; Erlinghagen & Knuth, 2004; Winkelmann & Zimmermann, 1998), as well as some recent studies (Giannelli, Jaenichen, & Rothe, 2013), shows that employment relations are largely stable over time and across different labour market

⁎

entry cohorts. However, an increasing number of findings point to a destabilisation of employment relations (Blossfeld et al., 2006; Buchholz, 2008; Struck, Grotheer, Schröder, & Köhler, 2007). In this context, an erosion of internal labour markets is reported because internal job changes and promotions have been observed less frequently since the mid-1990s (Diewald & Sill, 1980; Giesecke & Heisig, 2011). Against this backdrop, life course research is interested not only in levels of change and overall trends in employment stability but also in the determinants and mechanisms influencing employment trajectories (Blossfeld, 1985; Manzoni, 2012). Well-established labour market theories have in common that they explain action and outcomes in the labour market by predominantly focussing on individual factors, especially qualifications (Berg, 1981; Kalleberg & Sørensen, 1979). The “new structuralism” (Baron & Bielby, 1980) and recent approaches of HRM systems (Hendry, 2003; Lepak, Hui, Yunhyung, & Harden, 2006) stress the significance of accounting for firm characteristics. According to these approaches, companies – depending on their size, economic success or type of production – work with fixed core workforces or unprotected and unstable workforces.

Corresponding author at: Chair of Labour Studies, Otto-Friedrich University of Bamberg, Feldkirchenstraße 21, 96052 Bamberg, Germany. E-mail addresses: [email protected] (M. Dütsch), [email protected] (F. Ganesch), [email protected] (O. Struck).

https://doi.org/10.1016/j.alcr.2019.03.002 Received 29 January 2018; Received in revised form 19 November 2018; Accepted 2 March 2019 Available online 13 March 2019 1040-2608/ © 2019 Elsevier Ltd. All rights reserved.

Advances in Life Course Research 40 (2019) 43–84

M. Dütsch, et al.

However, existing approaches do not address the impact of macrostructural factors at the regional level. Sociologists such as Coleman (1990) and Esser (1996) have pointed to the importance of the broader social context for individuals. Just as it is important to recognise that individuals act within specific contexts, this also applies to the labour market. To some extent, employees’ options for decisions and action depend on resources or limitations that vary among regions because of different economic situations, structures of the labour force, etc. Thus, contextual factors must be taken into account, which leads to a multilevel explanation of social action in the labour market. Furthermore, theories of discrimination emphasise that in the labour market social actors tend to discriminate against others under certain conditions (Arrow, 1971; Phelps, 1972). Such behaviour is also contingent on structural frame conditions, e.g. when employers can make more use of observable characteristics of job applicants in dense regions where there is a high labour supply. Thus, it is essential to examine whether different groups of workers unequally benefit or suffer from specific regional settings (Hirsch, König, & Möller, 2013). Against this backdrop, this study raises the following question: to what extent do regional characteristics influence employment trajectories. Additionally, this study deepens the inquiry by investigating the possibly diverse effects of regional characteristics on different sociodemographic groups. This paper first (Section 2) reviews the current state of research. In Section 3, theoretical considerations are presented and hypotheses derived. Section 4 describes the datasets and explains the methodology and the variables. Empirical results on employment trajectories in Germany are presented in Section 4. A discussion of the results follows in Section 5, and the last Section concludes.

opportunities of employment trajectories, there are only a few studies that directly integrate regional indicators to explain individual employment trajectories from a life course perspective. Grotheer, Struck, Bellmann, and Gewiese (2004) analyse the impact of economic conditions on the employment stability of workers who have recently joined a company. They show that cycles of production and demand, as well as regional unemployment rates, stabilise employment relations in the structurally weaker regions of East Germany but destabilise them in West Germany. The authors explain this result based on the higher willingness to compromise of employees in structurally weaker areas in East Germany. Furthermore, the business cycle and the unemployment rate lead to more frequent transitions between firms in West Germany, as well as to lower job-to-job transitions and more transitions into unemployment in East Germany. Boockmann and Steffes (2005) found similar results when taking into account the previous year's unemployment rates at the federal state level. While they could not identify a clear effect of the unemployment rate on employment stability in West Germany, they did find a stabilising effect in East Germany. The likelihood of accomplishing inter-firm changes is lower for women in East Germany. In a further study on male employees, however, Boockmann and Steffes (2010) determined no significant effect on employment stability of the previous year's unemployment rates at the federal state level, but they did find a reduced inter-firm transition rate in West Germany. This research overview has shown that individual and firm-specific determinants influencing employment trajectories have been widely explored. Regional explanatory factors have been largely neglected or, at most, modelled by using regional unemployment rates. However, studies in regional economics mostly investigate macro-variables such as employment stocks and growth. Such research cannot be interpreted in terms of life course research due to the lack of a micro-foundation. Nevertheless, those studies hint at noteworthy parameters, especially concerning the extent of regional agglomeration and the regional endowment of human capital. These determinants will be examined systematically below by further focussing on diverse impacts on specific groups of workers.

2. State of research International research in regional economics has widely explored the association between agglomeration, productivity, and wage premiums. Combes, Duranton, and Gobillon (2008) and D’Costa and Overman (2014) showed that the sorting of workers with higher abilities in agglomerated areas explains higher wages in denser areas. Additionally, studies have indicated better matching, specialisation and learning lead to wage premiums in agglomerations (Combes et al., 2008; De La Roca & Puga, 2017; Duranton & Puga, 2004, chap. 48; Melo & Graham, 2014; Wheeler, 2008; for an overview see Combes & Gobillon, 2015, chap. 5). Furthermore, higher wages were found to be related to greater opportunities of individual accumulation of knowledge (Andersson, Klaesson, & Larsson, 2014; D’Costa & Overman, 2014; De La Roca & Puga, 2017; Matano & Naticchioni, 2016). According to Wheeler (2001), the local market size is positively associated with average productivity and greater between-skill-group wage inequality. Several studies have revealed that a large share of highly skilled workers in an area increases subsequent employment growth (Glaeser & Saiz, 2004; Shapiro, 2006; Simon & Nardinelli, 2002). Regional research on Germany points to stronger labour demand in densely populated areas in Germany (Blien, Kirchhof, & Ludewig, 2006; Dauth, 2013; Farhauer & Granato, 2006). Kelle (2016) found that population density is positively associated with wage growth rates. Furthermore, the regional skill structure positively influences employment growth. This finding can be explained by the share of the skilled workforce in an area (Blien & Wolf, 2002). Other studies have shown that the presence of a large proportion of highly skilled workers promotes employment growth in a region (Poelhekke, 2013; Schlitte, 2012; Shapiro, 2006; Südekum, 2008). However, Gerlach, Meyer, and Tsertsvadze (2002), Schlitte (2012) and Stephan (2001) found that skill segregation increases due to the presence of a high proportion of qualified employees in a region, leading to divergent development in terms of employment and wages. In addition to these studies on the effects of regional determinants on aggregated employment, which – at least – indicate the risks and

3. Theoretical considerations Individuals in modern societies pass through several stages during their life course (Kohli, 1985; Levy & Bühlmann, 2016; Mayer, 2009). They participate in education and then in family, as well as in the employment system. Finally, they leave the employment system and enter retirement. The employment period is considered a central stage of the life course (Kohli, 1985; Mayer, 2009) because it is mainly paid work that structures opportunities and risks during the life course. Against the backdrop that such opportunities and risks are the results of a cumulative process (Blossfeld, 1985; Manzoni, 2012), labour market and life course research has emphasised the disadvantages for lowskilled workers and females, as well as for foreign employees and workers with discontinuous employment trajectories. However, most of this research only focuses on individual factors (Bergemann & Mertens, 2004; Giesecke & Heisig, 2011) but ignore the fact that individual action always takes place within certain frame conditions (Coleman, 1990; Esser, 1996). Thus, individual employment trajectories can be affected by opportunities or limitations that arise from certain regional frame conditions. The significance of regional factors on economic parameters has recently been highlighted by the “new economic geography” (Krugman, 1991). As a starting point, Krugman (1991) relied on the work of Hirschman (1958) and developed a core-periphery model of economic activities. Thus, a range of divergent centripetal and centrifugal forces have to be taken into account. This approach looks at the effect of positive external factors and points to the mutual relationship among economies of scale, transportation costs, and migration. Centripetal forces lead to urbanisation effects because they encourage the concentration of economic activities 44

Advances in Life Course Research 40 (2019) 43–84

M. Dütsch, et al.

within a certain geographical area. Industrial centres are strengthened because firms and employees capitalise on agglomeration advantages. In the case of high economies of scale, a company tries to limit production to one single facility and to serve the market from there. To reduce transportation costs, the company will set up in a location with high population density and, therefore, higher demand. Both the workforce and firms are attracted to that regional economy in order to realise the agglomeration advantages created by a larger potential sales market and employee pool (Krugman, 1991, 1998). Hence, lower transportation costs and higher economies of scale increase the likelihood of the development of economic centres and peripheries alike. Referring to search and matching theories, Wheeler (2006) argues that greater pooling in the labour market should improve the match between employees and employers because urban density makes it “easier for workers to find the best jobs for themselves” (Glaeser & Maré, 2001: 322). Matching advantages arise if the chances of matching and the quality of successful matches between job seekers and vacancies increase with the size of the local labour market. As emphasised by Wheeler (2006), agglomeration benefits may arise due to improved firm-worker matches, which reflect the benefits of searching for jobs in urban rather than in non-urban labour markets. In urban areas, the number of job openings is larger and the search costs are lower. Accordingly, wage growth in cities should predominantly be related to job changes. Greater pooling occurs in denser regions; therefore, in agglomerated areas, the comparatively better matches between employers and employees due to a complementarity between worker skills and the firm's production processes lead to greater job stability. Additionally, agglomeration benefits are supposed to generate a wage increase (Glaeser & Maré, 2001). In the case of inter-firm job-to-job mobility in agglomerated areas, better matching between workers and employers provides wage increases as workers earn closer to their marginal product. Search and matching theories lead to the hypotheses that all employees working in core areas benefit from agglomeration advantages through high job stability (H1). Additionally, lateral and upward inter-firm mobility should take place to a greater extent in agglomerated regions (H2). By contrast, labour market theories also suggest that certain groups of employees can profit or suffer from particular regional frame conditions. This especially applies to those parts of the workforce often mentioned as being discriminated against in the labour market. This becomes evident when search and matching theories (Jovanovic, 1979; Wheeler, 2006) are combined with approaches of statistical discrimination (Arrow, 1971; Phelps, 1972). Employers who want to fill vacant positions will screen job applicants. In agglomerated areas, the larger pool of workers and job seekers leads to strong competition and enables an employer to select the subjectively most appropriate candidate. Discrimination theories emphasise that firms have limited information about the skills of job applicants. Therefore, employers use easily observable characteristics such as race or gender to infer the expected productivity of applicants. Observable characteristics function as a ‘screening device’ when recruiting staff in order to select the best potential employees. This reliance could lead to processes of segregation when employers anticipate lower productivity from specific groups of persons; for example, family formation and child-rearing often lead to interruptions in employment among women. This slows the accumulation of human capital through work experience and can contribute to the depreciation of previously accumulated human capital (Evertsson & Grunow, 2012; Mincer & Polachek, 1974). These mechanisms, or the anticipation thereof, can additionally affect employers’ evaluations of female workers’ productivity and lead to statistical discrimination against women because they are often expected to be less productive than males (Arrow, 1971; Phelps, 1972). Lower productivity is also often ascribed to foreign persons (Arrow, 1971; Phelps, 1972) and those who have experienced unemployment due to an assumed depreciation of previously accumulated human capital (Arulampalam, 2001; Manzoni & Mooi-Reci, 2011) or negative signalling effects of unemployment (Spence, 1973, 2002). Furthermore, Buch, Hamann,

Niebuhr, and Rossen (2017) show that in Germany, the share of workers who have acquired skills and qualifications, as well as professional degrees, increased in denser regions between 2000 and 2010. The gap between cities and rural areas in terms of the share of highly educated workers is significant. Additionally, according to Wheeler (2001), a bigger market size induces a greater degree of stratification of workers and firms by quality. In denser regions, greater sorting takes place so that higher quality firms tend to be matched with higher quality workers, leaving lower quality firms to match with lower quality workers. This leads to divers employment prospects and wage inequality across skill groups (Wheeler, 2001, 2006). This expectation can be hypothesised as follows: In agglomerated areas, the probability of experiencing lateral or downward job-to-job mobility is higher for women than for men (H3). This also applies to foreigners more often than to Germans (H4). Employees who have experienced longer unemployment periods during their employment trajectories face a greater risk of downward job-tojob mobility or unemployment (H5). Highly qualified workers should experience more stable employment in denser regions. Additionally, the lateral or upward job-to-job mobility of highly skilled workers is expected to be observed more often (H6). Another approach of regional research, the endogenous growth theory refers to the human capital theory (Becker, 1962, 1975) and points to a connection between the regional qualification structure and regional growth (Lucas, 1988; Romer, 1990). Thus, an increasing stock of qualifications and knowledge in a region is associated with rising worker productivity. Due to spillover effects stemming from formal and informal interactions between individuals, a high level of human capital endowment is considered an “engine of growth” for a region, even in case of a constant state of technology. On the one hand, this is attributed to the proximity of highly qualified and low-skilled workers, which may increase the likelihood of the latter learning from the former (Glaeser, 1999; Jovanovic & Rob, 1989). On the other hand, according to the theoretical assumption of Acemoglu (1996), companies with a large proportion of highly skilled workers will tend to invest in new production techniques, which in turn increase the productivity of the whole workforce. A further theoretical concept emphasises a complementary relation between different skills in the production process (Schlitte, 2012); thus, a large proportion of highly skilled workers also positively impacts the wages and employment prospects of less skilled workers. It can therefore be concluded that all qualification groups in regions with a high human capital endowment benefit from stable employment (H7) and good promotion prospects (H8). However, there are theoretical approaches that explain increasing inequalities in employment as the result of skill segregation (Acemoglu, 1999; Duranton, 2004). The general assumption of these theories is that workers will be deployed to an increasing degree, and dependent on their qualifications in the production system, as a result of technological and organisational change (Acemoglu, 2002; Bresnahan, Brynjolfsson, & Hitt, 2002). This leads to a rise in the productivity of the highly skilled workforce, while the productivity of the less-skilled workforce decreases (Acemoglu, 2002; Bresnahan et al., 2002). There exist several explanations for such skill segregation. According to Acemoglu (1999), employers orient their production technology towards the internal qualification structure. Against this background, organisations would more frequently invest in modern production technology, which would correspond particularly to the knowledge and skills of highly skilled workers and exclude the less qualified. This situation serves to increase the productivity and wages of only the highly skilled workforce. Similarly, Duranton (2004) assumes that human capital and technical equipment are complements. He argues that a rise in the supply of high-skill workers makes the production of higherquality goods more attractive, while some unskilled workers are excluded from the production system. This, in turn, leads to improved quality of the manufactured goods, and the demand on modern production processes as well as qualified employees will therefore increase further. Regarding this production system, the comparatively more 45

Advances in Life Course Research 40 (2019) 43–84

M. Dütsch, et al.

productive highly skilled workers can realise wage increases and experience more stable employment than the less qualified. The two approaches of Acemoglu (1999) and Duranton (2004) explain that segregation processes take place to different degrees depending on the qualifications of the labour force available for them. Thus, the level of regional human capital endowment of the labour force causes inequalities in employment prospects and wages between high- and lowskilled workers. Approaches based on skill segregation lead to the hypotheses that a high local level of human capital results in less stable jobs for lower-skilled employees (H9), while the highly skilled will profit through stable jobs or even upward inter-firm mobility (H10). By linking the theoretical approaches of skill segregation (Acemoglu, 1999; Duranton, 2004) to statistical discrimination theories (Arrow, 1971; Phelps, 1972), it can be assumed that employers in regions with a high regional endowment of human capital are more discerning in their employee selection. A large share of highly qualified employees in a region enables employers to choose potential employees who they expect to be comparatively more productive than their rival applicants. Accordingly, employers can rely on observable characteristics of applicants that they assume to be indicators of productivity. As already mentioned, females, foreigners and those who have experienced unemployment are often regarded as being less productive (Arrow, 1971; Arulampalam, 2001; Evertsson & Grunow, 2012; Phelps, 1972). The following can therefore be assumed: A high human capital endowment in an area increases the probability that female employees will experience lateral or downward job-to-job mobility (H11). The higher the local level of human capital, the greater the risks facing foreign workers (H12) and those who have experienced longer periods of unemployment during their employment trajectories (H13). The next section describes the data and the estimation strategy used to test the derived hypotheses.

periods by using assumptions because information is recorded only for periods in which a person receives unemployment benefits from the German Federal Employment Agency. As a result, the data do not account for those unemployed people who were not officially registered. Thus, a cleansing procedure is used to generate the three labour market states detailed above. An unemployment period is defined as a period in which the job seeker receives unemployment benefits for at least one day within a 90-day period. New employment or a job change is defined as a period of unemployment not exceeding 90 days. It is likely that in most of these cases, workers already knew their new employers when the previous job ended. The state “employment gap” is activated when no change of employment has occurred and unemployment benefits have not been received within the 90-day period.1 Thus, complete employment biographies can be constructed for those employees covered by the LIAB. Finally, we use two different data sources to obtain information on 96 German spatial planning regions, which are considered adequate for analysing regional labour markets (Rendtel & Schwarze, 1996; Schwarze, 1995). First, the Federal Institute for Research on Building, Urban Affairs and Spatial Development (BBSR) provided us with data on the types of regions.2 We distinguish among rural areas, regions undergoing the urbanisation process and urban areas. Second, the Establishment History Panel (BHP) is a 50 percent sample of all establishments throughout Germany with at least one employee subject to social security as of 30 June of a given year (Gruhl, Schmucker, & Seth, 2012). We calculated – based on the structure of employees by educational and vocational qualifications – the share of low-skilled, qualified, and highly qualified employees in the planning regions. These characteristics of the spatial planning regions are separately computed for each year to account for changes in the composition of the regional indicators. For example, according to the Harris-Todaro model, promising labour market conditions in a region may attract workers from other areas and increase the competition for jobs (Harris & Todaro, 1970); thus, job stability of the already employed may be influenced to some extent. Additionally, the interaction effects of region-specific factors and worker characteristics can be affected, as different groups of employees have different costs of mobility.3 This generated dataset permits simultaneous analyses of employers, employees, and the regional context. Employees and trainees are included unless they are exempted from social security. This, in turn, means that civil servants and the self-employed are not covered. Data are restricted to employees who started a new job between January 01, 2000, and December 31, 2010. Furthermore, employees whose income is above the income ceiling are excluded because this information is censored. Thus, the analysis sample consists of dependent employees aged 18–65 years. These criteria provide a sample of 1,138,538 workers, 6199 firms, and 96 regions. From a life course perspective, two different aspects of stability must be considered. The first is referred to as job stability and denotes the period when an employee works in a workplace in a specific company. The second is employment stability, which includes inter-firm job changes and comprises the duration of employment regardless of changes in employers. The latter is demanding due to the risk of becoming unemployed or of taking an inadequate position in case of

4. Data and method To investigate the influence of individual as well as contextual factors on employment trajectories, information about individuals’ employment states as well as wages must be provided in a longitudinal way. Additionally, as mentioned in the overview of current research, firm-specific determinants should be incorporated in research on employment careers. We use the Linked Employer–Employee dataset (LIAB) of the Institute for Employment Research, which allows controlling for factors at the individual, firm, and regional levels (Klosterhuber, Heining, & Seth, 2013). The LIAB combines administrative data on employees’ work history with a representative annual survey of 16,000 business establishments (IAB Establishment Panel). This study uses the “LIAB longitudinal Model 1993–2010” (LIAB LM 9310), which includes firms that took part in the survey continuously between 2000 and 2010. The employment and welfare recipient histories for the period from 1993 to 2010 were drawn from those persons who were employed in any of the LIAB firms for at least one day between 2000 and 2009. Data on employees are obtained from two different sources. First, the “employee history” contains data on individual employment history records submitted by employers to the German public pension insurance system. Thus, information on employment periods and wages is highly valid because failure to supply accurate information is a legal misdemeanour that may even result in a summary offence. One exception is individual information on the education variable; this variable is adjusted by imputation (Fitzenberger, Osikominu, & Völter, 2006). Second, the “benefit recipient history” is data on the receipt of unemployment benefits, unemployment assistance, or maintenance allowance. The employment histories are left-censored and can thus be tracked from January 1, 1993 onward. Administrative data on employees enable us to identify the three labour market states of “employment”, “unemployment”, and “employment gap”. However, we must sometimes construct unemployment

1 The state “employment gap", for example, consists of periods in military or civilian service, self-employment, maternity leave or in the hidden reserve. 2 This indicator is constructed by using four criteria: the population share in large and medium cities, the presence and size of a big city, the population density of the planning region, and the population density of the planning region without consideration of the big cities (BBSR, 2018). 3 In our empirical analyses, we do not directly separate the regional patterns of job-to job-mobility and its determinants from interregional mobility. However, the results of e.g. skill-specific interregional mobility caused by the different costs of mobility for qualification groups are represented by the regional structural features, which are computed separately for each year.

46

Advances in Life Course Research 40 (2019) 43–84

M. Dütsch, et al.

direct firm-to-firm transitions. These patterns must be considered in the estimation strategy. In order to assess job stability, we estimate a proportional hazard model (Blossfeld & Rohwer, 2002; Kalbfleisch & Prentice, 2002: 95ff.). For individual i employed in firm j in region k at time t, the hazard rate is specified as: ijk (t ;

x) =

0 (t )

exp[Zijk (t ) ]

assumption by stratifying the sample in some of the estimations for the planning regions (Kalbfleisch & Prentice, 2002: 118f.). This means that we have assumed a separate baseline hazard for each region while assuming the coefficients of the covariates to be the same. This is similar to the within-groups estimator in linear regression. In the stratified case, we can identify the coefficients of time-varying region-specific variables but not those of time constant variables varying only by stratum (i.e., region). A further assumption of the Cox proportional hazards model and the competing risks model regards independent censoring or independent risks, conditional on the covariates in the analysis model (Cox, 1972; Kalbfleisch & Prentice, 2002). Although this assumption is untestable, we try to account for it by considering a wide range of observables, especially lagged employment histories.

(1)

Compared to other parametric event history approaches, the proportional hazard model is considered flexible because the baseline hazard λ0(t) is assumed to be shifted proportionately by the covariates Zijk(t). The dependent variable in the following is the conditional hazard rate, which is defined as the instantaneous probability of exit from the current job. To estimate models for transitions to unemployment and a new job with the mobility pattern “lateral mobility”, “downward mobility” (defined as a decrease in wages of more than 5 percent), “upward mobility” (defined as an increase in wages of at least 10 percent), “transition into full-time job”, “transition into part-time job” and “transition into marginal employment”, we use the semi-parametric Cox partial likelihood estimator (Blossfeld & Rohwer, 2002; Cox, 1972; Kalbfleisch & Prentice, 2002: 99ff.).4 Transitions are treated as competing risks with distinct destination states (denoted by m): m ijk (t )

=

m 0 (t )

exp[z ijk (t )

m]

5. Findings 5.1. Descriptives In a first step in Fig. 1 we descriptively investigate regional heterogeneities that are of interest in the following analyses. This includes, on the one hand, population density. In Germany, there are 24 urban planning regions, 35 areas undergoing the urbanisation process and 37 rural areas. The last category can be observed in the eastern part of Germany to a greater extent than in the western part. On the other hand, we investigate the effect of regional human capital endowment. The share of highly qualified employees varies between 2.95 and 16.10 percent across the planning regions. Both measures indicate that there are considerable heterogeneities among German regions. In a second step, employment states and transitions upon leaving employment are examined by type of region and regional human capital endowment in order to obtain an indication of mobility patterns based on the frame conditions (Table 1). In the total sample, 22.7 percent of employment periods are right-censored. These periods persist throughout the observation period until the end of 2010. After leaving employment, 4 percent realise lateral, 4.1 percent downward and 7.1 percent upward job-to-job mobility without changing their type of employment; 2.9 percent of workers transition to full-time employment, 1.8 percent to part-time employment and 2.8 to marginal employment; 25.5 percent of employees become unemployed, and 27.6 percent transition into an employment gap.6 These transition patterns vary significantly between types of region. In urban areas more right-censored periods as well as lateral, upward and even downward mobility can be observed than in rural areas. Only 18.2 percent of workers change into unemployment in urban areas, while this share is considerably higher in areas in the process of urbanisation (24.7 percent) and rural areas (35.4 percent). Opportunities and risks during the employment history also seem to depend on the regional human capital endowment. We divided the planning regions into two groups based on the median share of highly qualified workers. Thus, there are fewer right-censored periods in areas where the share of highly qualified workers is above the median. In these regions that have a high human capital endowment, on the one hand, more lateral and upward job-to-job transitions take place; on the other hand, 28.1 percent of workers transition into unemployment. This risk is significantly lower in regions with a comparatively low human capital endowment. These descriptive findings indicate that working conditions differ systematically among German regions. Against this backdrop, we investigate these diverging transitions in more detail. We explain the effects of structural factors (type of region and regional human capital

(2)

In this case, the hazard function is specific for each destination state, such that separate parameter vectors are estimated for each state (Kalbfleisch & Prentice, 2002: 251ff.). To account for the hierarchical data structure, cluster-robust standard errors are calculated at the level of the regional planning regions. With 96 regional planning regions, the number of clusters can be considered sufficiently large to obtain unbiased standard errors (Angrist & Pischke, 2009; Kézdi, 2004). We use a wide range of person-, firm- and region-specific covariates on the determinants of employment trajectories.5 We estimate a model with individual characteristics only and a model with individual and firm characteristics, as well as a model that also contains region-specific factors. The full model contains demographic information (sex, age, level of education, nationality), information on the actual employment state, and information on the previous employment trajectory. Regarding the last, we condition on the entire employment history, which can be observed back to the year 1993. Firm-specific information concerns the reported employment prospects for the forthcoming year, the expected development of business volume in the current year compared to the previous year, vocational training, the share of highly qualified employees, the average gross daily wage of full-time employees, collective bargaining arrangements, works councils, firm size, and industry affiliation. Regional information is based on types of regions (rural areas, areas in the process of urbanisation, urban areas) and the human capital endowment. Because the analysis period is based on the years 2000–2010 and because it can be assumed that mobility processes vary due to the economic cycle (Giesecke & Heisig, 2011; Struck et al., 2007), we control for the economic situation using annual dummies. A key assumption in the Cox model is that of proportional hazards, which means that hazard functions for two strata must be proportional over time (Kalbfleisch & Prentice, 2002). We account for this 4 We use asymmetric boundaries for upward and downward mobility because we assume that employees perceive even slight reductions in wages as a worsening of their situation. In contrast, workers may perceive slight increases in wages as being normal and matching inflation. Therefore, the boundary for upward mobility is higher than that for downward mobility. The exit state “employment gap” is not reported for purposes of clarity. These results are available from the authors on request. 5 Descriptive statistics of individual, firm-specific, and region-specific characteristics are reported in Table 6 in the appendix.

6

In additional calculations, we expanded the employment gap to 180 days. The number of transitions into an employment gap decreases approximately 9 percentage points. However, there are no systematic differences in transition rates with regard to the type of region and the regional human capital endowment depending on the extension of the gap. 47

Advances in Life Course Research 40 (2019) 43–84

M. Dütsch, et al.

Table 1 Observed states and transitions after leaving employment. Exit states

Right-censored Lateral mobility Downward mobility Upward mobility Transition to full-time employment Transition to part-time employment Transition to marginal employment Unemployment Employment gap Transition to training Number of observations Pearson chi-square testsa

Total sample

Type of region

n

Urban area

in %

Human capital endowment in a region Area in urbanisation process

Rural area

Share of highly qualified workers in area is below the median

Share of highly qualified workers in area is above the median

Nin %nin %nin %nin %nn

in %

n

in %

n

in %

n

in %

n

in %

258776 45473 47055 80671 32974

22.7 4.0 4.1 7.1 2.9

116164 20201 20286 34896 12317

26.9 4.7 4.7 8.1 2.9

76823 13027 15405 25042 12534

21.0 3.6 4.2 6.8 3.4

65789 12245 11364 20733 8123

19.3 3.6 3.3 6.1 2.4

74953 10643 13617 19214 9049

26.0 3.7 4.7 6.7 3.1

183823 34830 33438 61457 23925

21.6 4.1 3.9 7.2 2.8

20131

1.8

8217

1.9

6786

1.9

5128

1.5

5271

1.8

14860

1.7

31700

2.8

12383

2.9

10474

2.9

8843

2.6

8383

2.9

23317

2.7

289962 314803 16993 1138538 –/–

25.5 27.6 1.5 100 –/–

78635 121792 6371 431262

18.2 90313 24.7 28.2 109447 29.9 1.5 5910 1.6 100 365761 100 chi2 = 185.63, p = 0.000

121014 83564 4712 341515

35.4 24.5 1.4 100

50734 91443 4754 288061

17.6 239228 31.7 223360 1.7 12239 100 850477 chi2 = 81.89, p = 0.000

28.1 26.3 1.4 100

Analyses are based on all employment periods of persons who took up employment between 1.1.2000 and 31.12.2010. The period of observation was 1.1.2000 to 31.12.2010. a Pearson chi-square tests were performed to explore the differences between types of region as well as between areas with unequal human capital endowment. Source: LIAB (LM 9310), own calculations.

Fig. 1. Distribution of population density (type of region) and human capital endowment (share of highly qualified employees). Analyses are based on 96 planning regions. Source: INKAR and Establishment History Panel, own calculations.

endowment) on individual employment trajectories. Furthermore, we assess in which way and to what extent structural and individual factors are correlated to explain the different effects of frame work conditions on the employment histories of certain groups of employees.

presented in Table 2, where we focus on types of regions, and in Table 3, where coefficients on human capital endowment are depicted. To facilitate the quantitative interpretation of the effects, the tables display the hazard ratios. Because our focus is on regional effects and their interactions with worker characteristics, as well as on the distinction between different exit states, we only discuss the results for the individual covariates we use in the interaction term. Firm-specific coefficients are shown in the tables but will not be interpreted below. The two specifications in Tables 2 and 3 show consistent results for individual and firm-specific determinants with only minor differences

5.2. Individual determinants of job stability and employment trajectories after exiting a job The results of the Cox estimations for individual, firm-specific and region-specific factors influencing employment trajectories are 48

Advances in Life Course Research 40 (2019) 43–84

M. Dütsch, et al.

Table 2 Estimates for individual, firm-specific and region-specific factors (focus on types of regions) influencing employment trajectories. Independent variables

Exit from job

Region-specific factors Type of region (Ref.: rural area) Area in urbanisation process 1.020 Urban area 1.145* Unemployment rate 0.995

Lateral mobility

Downward mobility

Upward mobility

Transition to fulltime employment

Transition to parttime employment

Transition to marginal employment

Unemployment

1.025 1.204* 0.994

1.122 1.012 0.997

1.072 1.275** 0.982**

1.148** 1.369*** 0.998

1.067 1.335*** 0.988*

1.112 1.089 0.959***

1.016 0.915 1.015*

0.549*** 1.088*** 0.998*** 1.140**

2.231*** 1.117*** 0.998*** 1.238***

1.074* 0.974** 1.001*** 1.333***

0.788*** 1.010* 0.999** 1.114**

1.050

0.903**

0.757***

0.934*

0.652***

0.576***

1.632***

0.339***

1.014

0.753***

0.640***

0.709***

1.603* 1.988***

0.829 1.104

0.397*** 0.313***

0.659*** 0.787***

1.540*** 1.110 0.800***

–/– 1.225* 0.719***

1.032 –/– 0.602***

1.027 0.411*** 0.769***

0.473*** 1.977***

1.493 2.084

1.083 2.248

1.129 1.160

1.259

5.790***

9.310***

1.616*

0.938 0.868 0.542*

2.555 3.077* 1.344

2.896* 2.820* 0.887

4.378*** 1.336 0.822

1.028

0.910**

1.099**

1.060*

1.119*

1.028

0.992 1.073 0.937 0.919

0.948 1.064* 0.978 0.954

Individual factors Sex (1 = female) 0.914*** 1.033 0.996 0.876** Age (in years) 0.975*** 1.031*** 0.998 1.020*** Age (in years squared) 1.000*** 1.000*** 1.000*** 0.999*** Nationality (1 = foreign) 1.196*** 0.861*** 0.980 0.916 Highest education level (Ref.: at most secondary school and no vocational training) At most secondary school and 0.834*** 0.920 0.736*** 1.025 vocational training *** ** *** Advanced secondary school 1.328 0.671 0.676 0.808 and no vocational training Advanced secondary school 0.868*** 0.802*** 0.592*** 1.194** and vocational training Polytechnic degree 0.773*** 0.840** 0.452*** 1.392*** University degree 0.911** 0.840** 0.455*** 1.418*** Employment state (Ref.: full-time) Part-time 0.923** 0.590*** 0.724*** 0.403*** Marginal employment 0.755*** 1.466*** 4.526*** 0.347*** Daily wage (deflated) 0.727*** 1.473*** 1.715*** 0.618*** Previous employment trajectory (Ref.: previous vocational education) First employment 0.939 0.829 0.727 0.631** 1.247 1.324 1.035 Share of full-time 1.363*** employment *** *** Share of part-time 1.667 1.639 2.647 1.504* employment Share of unemployment 1.831*** 0.951 1.332 0.969 Share of nonemployment 1.566*** 1.091 1.390 1.137 Left-censored 1.035 0.777 0.740 0.742

Firm-specific factors Employment prospects for the forthcoming year (Ref: will be approximately constant) The level of employment will 1.014 1.056 1.136*** 1.159** 1.089* 1.122** increase The level of employment will 1.072** 1.216** 1.201*** 1.059 1.069 1.120* decrease Not sure at present 1.062* 1.106 0.986 1.022 0.979 0.992 Expected development of business volume in the current year compared to previous year (Ref.: it is expected to remain constant) * * It is expected to increase 0.974 1.103 1.019 1.097 1.038 0.953 1.126* 1.100* 1.084 1.025 1.077* It is expected to decrease 1.057** Do not know at present 0.986 1.371*** 0.966 0.971 0.874 0.925 1.008 1.081 0.976 1.013 1.082* Apprenticeships for vocational 0.926** training are offered (1 = yes) Share of high-qualified 0.913* 0.924 0.678*** 0.933 0.675*** 0.646*** employees Works council (1 = yes) 0.950* 0.914* 1.013 0.952 1.011 1.107 Average gross daily wage of 0.993*** 0.990*** 0.990*** 0.989*** 0.991*** 0.990*** full-time employees Collective agreement (Ref.: establishment not bound by a collective agreement) Sectoral collective agreement 0.998 1.017 1.062 1.103 0.931 1.004 Company collective 0.981 0.973 0.908 1.010 0.981 0.931 agreement * Orientation of sectoral 0.958 0.950 0.984 0.941 0.969 0.975 collective agreement Firm size (Ref.: small firm) Small-medium-sized firm 1.075*** 0.944 1.057 1.123 1.018 1.059 Medium-sized firm 1.034 0.771** 0.935 0.992 0.930 0.884 ** ** Larger firm 0.977 0.722 0.827 0.919 0.851 0.825* Sector (Ref.: manufacturing industry) Agriculture, forestry, and 1.302*** 1.406 1.212 1.029 1.170 1.460* mining Construction 1.731*** 1.726*** 1.428*** 1.113 1.162 1.312 Trade 1.191*** 1.722*** 1.123 1.396*** 1.304** 1.912*** 2.325*** 1.600*** 2.549*** 1.662*** 2.517*** Services for firms 1.402*** Other services 1.162*** 1.761*** 1.102 1.493*** 1.185* 2.242*** Year (Ref.: 2000) 2001 2002

1.073*** 1.144***

0.988 0.932

0.934 0.913

0.990 0.946

0.948 0.990

1.003 1.036

0.662*** *

0.977

0.909 0.990***

0.926 0.987***

1.005 1.070

1.017 0.927

0.922

0.982

1.051 1.055 1.033

0.929 0.834** 0.731***

1.329*

1.162

1.228* 1.116 1.388*** 1.000

2.215*** 1.057 1.303** 1.004

1.087 1.399***

1.145*** 1.298***

(continued on next page) 49

Advances in Life Course Research 40 (2019) 43–84

M. Dütsch, et al.

Table 2 (continued) Independent variables

Exit from job

Lateral mobility

Downward mobility

Upward mobility

Transition to fulltime employment

Transition to parttime employment

Transition to marginal employment

Unemployment

2003 2004 2005 2006 2007 2008 2009 2010

1.175*** 1.228*** 1.282*** 1.262*** 1.271*** 1.204*** 1.036 0.861***

1.006 1.066 1.059 1.040 0.985 0.887 0.667*** 0.418***

0.924 0.999 1.118 1.174* 1.119 1.002 0.833** 0.551***

0.930 1.098 1.438*** 1.509*** 1.470*** 1.191** 1.084 0.780**

1.126** 1.212*** 1.524*** 1.690*** 1.581*** 1.410*** 1.327*** 1.072

0.991 1.070 1.376*** 1.403*** 1.377*** 1.400*** 1.246*** 0.939

2.071*** 2.071*** 2.393*** 2.927*** 2.833*** 2.588*** 2.201*** 1.495***

1.344*** 1.155** 0.856* 0.785*** 0.889 0.869 0.783 1.034

Number of observations −2*LL (starting values) −2*LL (final estimates) AIC BIC

1138538 12018907 11862703 2599.020 2599.020

1138538 11544886 1141044 459.908 459.908

1138538 1198638 11760050 476.045 476.045

1138538 2031604 19780952 798.789 798.789

1138538 8416876 7392044 1096.773 1098.257

1138538 5112364 4763068 706.446 707.930

1138538 405090 381730 763.569 764.217

1138538 3707672 3579912 715.993 716.059

Analyses are based on all employment periods of persons who took up employment between 1.1.2000 and 31.12.2010. The period of observation was from 1.1.2000 to 31.12.2010. *p < 0.05; **p < 0.01; ***p < 0.001. Hazard ratios for the Cox partial-likelihood estimates; Calculation of cluster-robust standard errors for 96 planning regions. Source: LIAB (LM 9310), own calculations.

in coefficients and significances. Results for sex, age, and nationality are highly significant (see 2 or 3). Female employees leave their jobs less often than do men. However, in the case of a job exit, they comparatively seldom manage an upward transition or a transition to a full-time job. The probability of women filling a part-time job or of taking on marginal employment is increased.7 This finding indicates that women often take responsibility for child-rearing. However, females face a lower unemployment risk. The likelihood of leaving a job decreases with the age of the employee. In addition, lateral or upward mobility, as well as transitions into full-time and part-time jobs become more likely with increasing age. Changes in marginal employment become rarer as employees get older but regain importance at older ages. Nationality is still an important criterion in the German labour market: Foreign employees are employed less stably, and in the case of a job exit their unemployment risk is increased compared to Germans. Regarding the highest level of education, it becomes clear that employees with vocational training have more stable employment than persons who completed secondary school but not vocational training. In addition, employees with vocational training or university degrees experience downward mobility less often. They also have a lower risk of switching from full-time or part-time work to marginal employment compared to those without a vocational training qualification. Furthermore, they have a lower unemployment risk. This confirms the considerable importance of vocational education in Germany. Employees with a university degree achieve educational returns through wage increases after a job transition. Compared to full-time employees, part-time and marginal workers leave their jobs less frequently. Part-time and marginal workers are usually women who have reduced their working hours to arrange their job around their family responsibilities. This willingness or need to adapt results in lower unemployment risks for women. Additionally, the higher the daily wage, the lower the likelihood of leaving the company or becoming unemployed. In the case of inter-company mobility, higher earners are less likely to be upwardly and more likely to be downwardly mobile. Life course research emphasised the significance of endogenous causalities in employment trajectories. Our results show that job-to-job mobility can be more frequently observed among persons who, based on their previous employment history, were employed full-time or part-time proportionately longer or have been unemployed or in a social security

gap. Employees who were previously often employed full-time are more likely to return to full-time work, whereas those who were previously often employed part-time are less likely to move up and more often work part-time or in marginal jobs. As the individual's share of previous periods of unemployment increases, the risk of becoming unemployed increases. 5.3. Region-specific determinants of job stability and employment trajectories after exiting a job Regarding the impact of regional heterogeneities on individual employment trajectories, Dütsch and Struck (2014: 114) show, by applying a random-effects regression, that variance explained by the regional level is small compared to that explained by the firm- and person-level. However, our results in Table 7 in the appendix, which presents coefficient estimates of Cox regressions using different specifications, point to selection processes into specific regions.8 In Table 7, model 1 is taken from an estimation with individual-specific variables only, whereas the second column adds firm characteristics. The third and fourth columns also include region-specific factors without and with region-fixed effects. Comparing the results from specifications with and without firm characteristics, we can assess the importance of worker self-selection into certain firms (Mumford & Smith, 2004). Furthermore, comparing results from models with and without regional characteristics and region-fixed effects, we can assess the importance of the selection of firms and workers into certain regions. Against this backdrop, in Tables 2 and 3, we investigate coefficient estimates for region-specific factors affecting employment trajectories by referring to hypotheses derived from the theories presented in Section 3. Thus, employees working in urban areas have a higher probability to exit from a job compared to employees in rural areas (see Table 2). Furthermore, employees in urban areas are more likely to directly change their employer and to change from marginal employment to a full-time or part-time job. Almost no differences exist between areas undergoing the urbanisation process and rural areas, with the exception of the higher probability of entering a full-time job. Thus, one the one hand, we must reject hypothesis 1, namely that all employees working in core areas benefit from agglomeration advantages through 8

In Table 7, only model specifications for estimations where we focus on types of regions are shown. Analogously, we performed the same estimations using human capital endowment at the regional level. Because we obtained the same results, this table is not presented in the manuscript, but is available from the authors on request.

7

Marginal employment (“mini-jobs”) is a specific form of employment in which employees can earn €450 per month free of income tax and social security contributions. However, they receive no health insurance and only optional pension insurance. 50

Advances in Life Course Research 40 (2019) 43–84

M. Dütsch, et al.

Table 3 Estimates for individual, firm-specific and region-specific factors (focus on human capital endowment) influencing employment trajectories. Independent variables

Exit from job

Lateral mobility

Region-specific factors Human capital endowment (Ref.: share of low-skilled workers) Share of skilled workers 1.003 0.992 Share of highly skilled 1.009 1.012 workers Unemployment rate 0.995 0.995

Downward mobility

Upward mobility

Transition to fulltime employment

Transition to parttime employment

Transition to marginal employment

Unemployment

0.988 1.011

0.995 1.024

0.980* 1.022**

1.002 1.034**

1.020 1.019*

1.021* 1.024*

0.998

0.979*

0.998

0.980**

0.946***

1.002

0.546*** 1.088*** 0.998*** 1.143*

2.214*** 1.116*** 0.998*** 1.247***

1.071* 0.975*** 1.000** 1.342***

0.788*** 1.011* 1.000*** 1.103**

1.051

0.915**

0.747***

0.927*

0.673***

0.617***

1.638***

0.337***

1.024

0.768***

0.635***

0.701***

1.605* 2.035***

0.854 1.141

0.393*** 0.313***

0.648*** 0.777***

2.643*** 1.109 0.801***

–/– 1.215* 0.718***

1.028 –/– 0.602***

1.026 0.416*** 0.771***

0.477** 1.985** 1.279

1.518 2.112 5.886***

1.066 2.211 9.209***

1.113 1.023 2.039***

0.945 0.879 0.547*

2.606 3.115* 1.369

2.811* 2.785* 0.866

4.295*** 1.322 0.813

1.084*

1.096*

1.026

0.906**

1.062

1.105*

1.103**

1.058*

1.120*

1.034

0.998 1.074 0.944 0.927

0.947 1.064* 0.990 0.962

Individual factors 1.031 0.991 0.872** Sex (1 = female) 0.913*** Age (in years) 0.975*** 1.030*** 0.997 1.019** Age (in years squared) 1.000*** 1.000*** 1.000*** 1.000*** Nationality (1 = foreign) 1.196*** 0.863** 0.978 0.917 Highest education level (Ref.: at most secondary school and no vocational training) 0.933 0.748*** 1.037 At most secondary school and 0.835*** vocational training *** ** *** Advanced secondary school 1.340 0.691 0.716 0.834 and no vocational training Advanced secondary school 0.869*** 0.812*** 0.602*** 1.207*** and vocational training Polytechnic degree 0.774*** 0.862* 0.465*** 1.421*** University degree 0.914* 0.867* 0.475*** 1.457*** Employment state (Ref.: full-time) Part-time 0.924** 0.589*** 0.725*** 0.403*** 1.428*** 4.415*** 0.347*** Marginal employment 0.757*** Daily wage (deflated) 0.727*** 1.459*** 1.692*** 0.618*** Previous employment trajectory (Ref.: previous vocational education) First employment 0.938 0.841 0.742 0.635** Share of full-time employment 1.361*** 1.265 1.354 1.041 1.657 2.696*** 1.516* Share of part-time 1.667*** employment Share of unemployment 1.829*** 0.965 1.366 0.979 Share of nonemployment 1.567*** 1.098 1.409 1.144 Left-censored 1.034 0.784 0.751 0.747 Firm-specific factors Employment prospects for the forthcoming year (Ref: will be approximately constant) The level of employment will 1.010 1.043 1.118* 1.143** increase The level of employment will 1.069** 1.212** 1.192*** 1.055 decrease Not sure at present 1.061** 1.106 0.983 1.020 Expected development of business volume in the current year compared to previous year (Ref.: * It is expected to increase 0.974 1.102 1.021 1.095* 1.125* 1.099* 1.081 It is expected to decrease 1.058** Do not know at present 0.986 1.356*** 0.968 0.965 1.007 1.082 0.976 Apprenticeships for vocational 0.928** training are offered (1 = yes) Share of high-qualified 0.950* 0.945 0.704*** 0.947 employees * * Works council (1 = yes) 0.912 0.915 1.010 0.951 Average gross daily wage of full- 0.993*** 0.990*** 0.989*** 0.989*** time employees Collective agreement (Ref.: establishment not bound by a collective agreement) Sectoral collective agreement 1.006 1.027 1.081 1.120 Company collective 0.981 0.957 0.899 1.000 agreement Orientation of sectoral 0.961 0.950 0.993 0.944 collective agreement Firm size (Ref.: small firm) Small-medium-sized firm 1.072** 0.940 1.057 1.114 Medium-sized firm 1.035 0.783** 0.955 1.004 ** Larger firm 0.971 0.715 0.821 0.904 Sector (Ref.: manufacturing industry) *** Agriculture, forestry, and 1.308 1.431 1.240 1.051 mining Construction 1.721*** 1.681*** 1.386*** 1.081 Trade 1.184*** 1.693*** 1.094 1.369*** 2.245*** 1.530*** 2.461*** Services for firms 1.388*** Other services 1.155*** 1.708*** 1.071 1.451*** Year (Ref.: 2000) 2001 2002

1.067*** 1.135***

0.974 0.907

0.920 0.890*

0.972 0.916

0.976 0.979 it is expected to remain constant) 1.032 0.953 1.024 1.071 0.882 0.923 1.014 1.083 0.675***

0.659***

0.646*** *

0.963

1.000 0.991***

1.089 0.990***

0.906 0.990***

0.929 0.987***

0.949 0.971

1.021 0.919

1.020 1.073

1.035 0.938

0.974

0.979

0.932

0.995

1.014 0.932 0.854**

1.048 0.893 0.798**

1.046 1.044 1.019

0.926 0.827** 0.717***

1.188

1.484*

1.319*

1.148

1.153 1.306*** 1.635*** 1.191*

1.258 1.872*** 2.359*** 2.180***

1.221* 1.117 1.381*** 1.001

2.197*** 1.042 1.295*** 0.994

0.942 0.973

0.978 0.993

1.078 1.387***

1.144*** 1.297***

(continued on next page) 51

Advances in Life Course Research 40 (2019) 43–84

M. Dütsch, et al.

Table 3 (continued) Independent variables

Exit from job

Lateral mobility

Downward mobility

Upward mobility

Transition to fulltime employment

Transition to parttime employment

Transition to marginal employment

Unemployment

2003 2004 2005 2006 2007 2008 2009 2010

1.163*** 1.208*** 1.259*** 1.228*** 1.223*** 1.149*** 0.983 0.812***

0.960 0.990 0.964 0.928 0.867 0.769* 0.574*** 0.353***

0.884 0.928 1.019 1.051 0.989 0.870* 0.718*** 0.469***

0.884 1.016 1.303** 1.335*** 1.272* 1.004 0.904 0.640***

1.097* 1.179*** 1.464*** 1.613*** 1.498*** 1.321*** 1.232* 0.996

0.931 0.965 1.219** 1.200** 1.315*** 1.111 0.970 0.720**

2.068*** 2.375*** 2.850*** 2.865*** 2.689*** 2.401*** 2.018*** 1.348***

1.355*** 1.165** 0.868* 0.782*** 0.856 0.828 0.737** 0.965

Number of observations −2*LL (starting values) −2*LL (final estimates) AIC BIC

1138538 12018907 11862703 2599.020 2599.020

1138538 577244 570394 1140901 1141573

1138538 599318 587848 1175809 1176481

1138538 1015802 988805 1977722 1978395

1138538 420843 369622 739355 740015

1138538 255618 238099 476309 476969

1138538 405090 381732 763574 764235

1138538 3707672 3579788 7159688 7160360

Analyses are based on all employment periods of persons who took up employment between 1.1.2000 and 31.12.2010. The period of observation was from 1.1.2000 to 31.12.2010. *p < 0.05; **p < 0.01; ***p < 0.001. Hazard ratios for the Cox partial-likelihood estimates; Calculation of cluster-robust standard errors for 96 planning regions. Source: LIAB (LM 9310), own calculations.

upward mobility. In areas undergoing the urbanisation process, these workers comparatively seldom realise downward mobility, upward mobility or transitions between different forms of employment. Employees who completed a polytechnic degree benefit most from working in an urban area: they have a lower risk of lateral or downward job-tojob mobility, a greater likelihood of upward mobility, and lower risks of transitions into marginal employment or even unemployment. In areas undergoing the urbanisation process, these workers are employed with greater stability and face a lower risk of downward mobility in the case of a job exit. In denser regions, employees who have obtained a university degree have a lower probability of upward mobility and of transitioning into full-time or part-time jobs. They are more likely to benefit in rural areas, as the baseline coefficients show. These findings, especially regarding the vocationally trained and employees with a polytechnic degree, support hypothesis 6, namely that highly qualified workers should experience more stable employment and lateral or upward job-to-job mobility in denser regions. Results in Table 5 regarding the relationship between sex and the regional human capital endowment indicate that the employment stability of female workers increases as the share of skilled workers increases. However, the baseline coefficient reveals that the greater the share of low-skilled workers in a region, the lower their job stability is. In the case of job-to-job transitions, female workers can, on the one hand, more often realise changes in full-time jobs, but on the other hand, they are at a greater risk of marginal employment as the share of skilled workers increases. Thus, hypothesis 11, namely that female employees’ probability of experiencing lateral or downward job-to-job mobility is increased in areas with a high human capital endowment, must be rejected. Foreign workers obviously profit from a greater share of skilled employees in a region because, in the case of a job exit, their opportunities for upward mobility increase while their unemployment risks decrease. In contrast, according to the baseline coefficients, foreign workers face greater unemployment risks the higher the share of low-skilled workers there is in a region. These findings do not support hypothesis 12, which assumed a greater risk for foreign employees in areas with a high local level of human capital. As the regional endowment of human capital increases, a greater share of unemployment periods in a person's pervious employment trajectory leads to less upward mobility and fewer transitions into full-time employment; it also leads, to a lesser extent, into unemployment. However, their unemployment risks considerably increase the greater the share of lowskilled workers there is in a region. Again, the assumption of hypothesis 13, namely that risks increase, must be rejected. Employees who completed, at most, secondary school and vocational training are at a

high job stability. On the other hand, hypothesis 2 holds true, namely that both lateral and upward inter-firm mobility take place to a greater extent in agglomerated regions. Table 3 shows that the higher the share of skilled workers in a region, the higher the risk of becoming unemployed. This also applies to the regional endowment of highly skilled workers. Obviously, the pressure of competition is greater between employees in such regions. Thus, we must reject hypothesis 7, namely that employees in regions with a high endowment of human capital benefit from stable employment. This also applies to hypothesis 8 because of the higher unemployment risks. 5.4. Interaction effects between region-specific and individual factors on employment trajectories We further explore whether interaction effects exist between individual and region-specific determinants. To increase clarity, only results for baseline coefficients and interaction terms are displayed in Tables 4 and 5. However, the models contain all explanatory variables that are reported in Tables 2 and 3. Estimates in Table 4 for interaction terms between sex and types of region show that females are more likely than males to exit their jobs in urban areas and in areas undergoing the urbanisation process. Furthermore, they seem to benefit from working in urban areas because they are able to realise lateral or upward firm-to-firm mobility and are at a lower risk of filling a position in marginal employment. Accordingly, hypothesis 3, namely that the probability of experiencing lateral or downward job-to-job mobility is higher for women than for men, can be rejected. Regarding nationality, foreign workers are more stably employed in denser regions. Additionally, their probability of transitioning to marginal employment is higher, especially in urban areas. This finding contradicts the lateral or downward inter-firm job changes predicted in hypothesis 4. The results regarding previous employment trajectory show that in agglomerated areas, the risks of downward mobility or transitioning into unemployment increase as the share of previous unemployment periods increases. This is in line with hypothesis 5. Employees who completed, at most, secondary school and vocational training are more stably employed in urban areas. Additionally, their probability of transitioning into a different type of employment is lower in urban areas as well as in areas undergoing the urbanisation process. Employees with an advanced secondary school and vocational training certificates are less often faced with lateral or downward inter-firm mobility, or with transitions into full-time jobs in urban areas. However, they have a greater opportunity to realise 52

1.029 1.208* 0.964 0.977 0.981

1.045 0.973

1.331 1.054

0.681* 0.649** 1.010 1.049 1.819***

1.014 1.206* 0.776

1.049 1.310***

1.092** 1.178*** 1.035 1.071* 1.710**

1.002 1.066 0.904

0.980 0.972 0.836***

Exit from job Lateral mobility

53 0.632** 1.134 0.627***

1.339** 1.305** 0.764*

0.858 0.823 0.630**

0.733 0.703* 0.824** 0.729

0.746

*

0.771 0.712* 0.893

0.739 0.761 0.768** 0.914

0.688

*

0.667*

0.736*

0.876 2.053*** 0.733** 0.531*** 0.397***

1.240 1.322*

1.153 0.887

1.078 1.103 2.858*

1.221 1.509*

1.109 1.059 0.971

0.912 0.742***

1.170* 1.274** 1.242**

Transition to marginal employment

0.773 0.464*** 0.911

1.029 1.096 0.922 1.001

0.884

0.862

0.816

0.957 0.510*** 0.778** 0.718*** 0.749**

1.019 0.969

1.169* 1.371***

0.958 0.829** 3.832***

1.047 1.126

1.016 0.908 1.026

1.044 1.097

0.998 0.882* 0.758***

Unemployment

Analyses are based on all employment periods of persons who took up employment between 1.1.2000 and 31.12.2010. The period of observation was from 1.1.2000 to 31.12.2010. Models additionally contain all explanatory variables that are reported in Table 2. *p < 0.05; **p < 0.01; ***p < 0.001. Hazard ratios for the Cox partial-likelihood estimates; Calculation of cluster-robust standard errors for 96 planning regions. Source: LIAB (LM 9310), own calculations.

0.717 0.774* 0.748*** 0.644*

0.668 1.004 0.738** 1.145 0.767

0.703

**

0.667*

0.515** *

0.821*

0.797**

1.027

0.470***

1.058 0.901 0.882 0.997 1.544**

1.095 0.941 1.141 1.746*** 2.691***

1.313*** 1.584***

1.179 1.206

0.961 1.201 1.118 1.329** 1.784***

1.088 1.172

1.025 1.282*** 2.267

1.074 1.065

1.064 1.332*** 1.166

0.918 0.916

1.133 1.419*** 2.382***

Transition to part-time employment

1.417*** 1.777***

1.267 1.243

1.361* 1.451**

1.114 1.271*** 0.804

0.890 0.884

1.151* 1.375*** 1.075

0.998 0.877

1.151* 1.470*** 1.188**

Transition to full-time employment

1.116 1.174

1.019 1.219** 0.834

1.107 0.882

1.066 1.287** 0.966

1.016 1.296***

0.756** 1.282* 0.874**

Upward mobility

1.053 1.161 1.049

1.069 0.982

1.117 1.246** 0.973

1.087 1.136

1.081 1.174* 0.917

Downward mobility

Type of region (Ref.: rural area) Area in urbanisation process 1.124 1.226 1.284** Urban area 1.132* 1.243* 1.335** Highest education level (Ref.: at most secondary school and no vocational training) At most secondary school and vocational training 0.903* 0.972 0.780*** 1.436* 1.103 Advanced secondary school and no vocational training 1.509*** Advanced secondary school and vocational training 0.882* 0.9278 0.661*** *** Polytechnic degree 0.807 0.873 0.579*** University degree 0.972 0.978 0.626*** Highest education level (Ref.: at most secondary school and no vocational training) × types of region (Ref.: rural area) At most secondary school and vocational training × area 0.897 0.595** 0.805 undergoing urbanisation process Advanced secondary school and no vocational training × area 0.831 1.062 0.769* undergoing urbanisation process 0.915 0.797 0.649** Advanced secondary school and vocational training × area undergoing urbanisation process 0.787* 0.633*** Polytechnic degree × area undergoing urbanisation process 0.857* University degree × area undergoing urbanisation process 0.912 0.878 0.747 At most secondary school and vocational training × Urban area 0.895* 1.025 0.975 Advanced secondary school and no vocational training × Urban 0.878 0.813 0.888 area ** Advanced secondary school and vocational training × Urban area 1.012 0.506 0.631** 0.527* Polytechnic degree × Urban area 1.001 0.341*** University degree × Urban area 0.919 0.978 0.919

Type of region (Ref.: rural area) Area in urbanisation process Urban area Share of unemployment periods Share of unemployment periods × types of region (Ref.: rural area) Share of unemployment periods × area in urbanisation process Share of unemployment periods × urban area

Types of region (Ref.: rural area) Area in urbanisation process Urban area Nationality (1 = foreign) Nationality (1 = foreign) × types of region (Ref.: rural area) Foreign × area in urbanisation process Foreign × urban area

Type of region (Ref.: rural area) Area in urbanisation process Urban area Sex (1 = female) Sex (1 = female) × types of region (Ref.: rural area) Female × area in urbanisation process Female × urban area

Independent variables

Table 4 Estimates for cross-level interactions between types of regions and individual factors influencing employment trajectories.

M. Dütsch, et al.

Advances in Life Course Research 40 (2019) 43–84

1.003 0.994 1.008 1.020 0.969 2.178 share of low-skilled workers) 1.001 0.989 1.011 0.943

Human capital endowment (Ref.: share of low-skilled workers) Share of skilled workers Share of highly skilled workers Share of unemployment periods Share of unemployment periods × human capital endowment (Ref.: Share of unemployment periods × share of skilled workers Share of unemployment periods × share of highly skilled workers 1.002 1.032* 1.005 0.959*** 0.953***

0.952*** 0.960

0.939*** 0.958*

0.989 1.027*** 1.264

0.990 0.992

1.027** 1.002

1.010 0.999 0.997 1.017 1.958

0.981* 1.022** 0.481

54

0.045*** 0.717 0.006*** 0.002*** 0.065* 1.034*** 0.981

1.069*** 1.029 1.040** 1.061

0.019*** 6.598*** 0.335 0.001*** 0.005*** 1.051*** 0.959** 1.012 1.083*** 1.063*** 1.031* 0.942* 0.994 1.018 1.027

0.331 1.770 0.447 0.005*** 0.197 1.017 0.971 1.009 1.061*** 1.033 0.984 1.026 1.028 1.082** 0.974

Analyses are based on all employment periods of persons who took up employment between 1.1.2000 and 31.12.2010. The period of observation was from 1.1.2000 to 31.12.2010. Models additionally contain all explanatory variables that are reported in Table 3. *p < 0.05; **p < 0.01; ***p < 0.001. Hazard ratios for the Cox partial-likelihood estimates; Calculation of cluster-robust standard errors for 96 planning regions. Source: LIAB (LM 9310), own calculations.

1.010 0.810**

1.055**

1.052***

0.994 0.994

0.920*** 0.968*

1.042*** 1.029* 5.105***

0.969*** 0.986

1.023* 1.024* 3.123***

0.993 1.008

1.023* 1.020 1.441

Unemployment

0.984 1.001

1.021 1.029

1.018 1.015 0.377

0.972 1.001

1.020 1.016 1.890

1.021* 0.994

1.008 1.022* 0.363

Transition to marginal employment

0.988 1.040**

0.998 0.9998

1.004 1.035 1.935

1.012 1.037*

1.001 1.030** 0.342

0.992 0.971

1.042*** 1.017

0.994 1.024 0.163**

1.006 1.053*** 6.230***

Transition to part-time employment

0.960*** 1.014 0.018***

Transition to full-time employment

0.987 1.010 0.467

0.984 1.004

1.003 1.023 2.149

Upward mobility

Human capital endowment (Ref.: share of low-skilled workers) 0.990 0.971** 0.974 0.941*** Share of skilled workers 0.987* Share of highly skilled workers 0.997 0.991 0.992 0.994 1.006 Highest education level (Ref.: at most secondary school and no vocational training) At most secondary school and vocational training 0.098*** 0.850 0.268* 0.234 0.014*** Advanced secondary school and no vocational training 4.758 0.317 0.067* 0.181 0.222 0.870 0.038*** 0.207 0.007** Advanced secondary school and vocational training 0.259* *** *** Polytechnic degree 0.161 0.580 0.008 0.292 0.006** University degree 0.470 0.143 0.005*** 0.051** 0.611 Highest education level (Ref.: at most secondary school and no vocational training) × human capital endowment (Ref.: share of low-skilled workers) 0.998 1.010 1.018 1.056*** At most secondary school and vocational training × share of 1.026*** skilled workers Advanced secondary school and no vocational training × share 0.989 1.008 1.029 1.021 1.018 of skilled workers * Advanced secondary school and vocational training × share of 1.013 0.993 1.031 1.019 1.065*** skilled workers Polytechnic degree × share of skilled workers 1.016** 1.001 1.045** 1.021 1.051** University degree × share of skilled workers 1.011 1.026 1.063*** 1.047** 1.029* At most secondary school and vocational training × share of 1.019*** 1.021 1.011 1.029 1.021 highly skilled workers Advanced secondary school and no vocational training × share 1.001 1.010 1.027 1.024 1.028 of highly skilled workers ** ** * ** Advanced secondary school and vocational training × share of 1.023 1.060 1.049 1.066 1.026 highly skilled workers Polytechnic degree × share of highly skilled workers 1.020* 1.033 1.009 1.011 1.080* 0.990 0.999 1.030 0.983 University degree × share of highly skilled workers 0.912**

1.002 0.992 1.011* 1.012 0.519 0.552 of low-skilled workers) 1.013 1.009 0.981 0.989

0.988 0.999

0.971 0.997

Human capital endowment (Ref.: share of low-skilled workers) Share of skilled workers Share of highly skilled workers Nationality (1 = foreign) Nationality (1 = foreign) × human capital endowment (Ref.: share Foreign × share of skilled workers Foreign × share of highly skilled workers

0.993 1.011 2.648

Downward mobility

1.006 1.013 5.217**

Exit from job Lateral mobility

Human capital endowment (Ref.: share of low-skilled workers) Share of skilled workers 1.010 Share of highly skilled workers 1.010 Sex (1 = female) 2.135* Sex (1 = female) × human capital endowment (Ref.: share of low-skilled workers) Female × share of skilled workers 0.988** Female × share of highly skilled workers 1.004

Independent variables

Table 5 Estimates for cross-level interactions between regional human capital endowment and individual factors influencing employment trajectories.

M. Dütsch, et al.

Advances in Life Course Research 40 (2019) 43–84

Advances in Life Course Research 40 (2019) 43–84

M. Dütsch, et al.

disadvantage as the regional share of high or highly skilled workers increases. Those employees experience less stable employment and have a greater risk of entering into marginal employment or even unemployment. Employees who completed advanced secondary school but no vocational training are at a lower risk of transitioning into marginal employment as the share of high or highly skilled workers increases. Workers possessing an advanced secondary school and vocational training certificate have a higher probability of less stable employment as the regional share of highly skilled workers increases. Furthermore, their chance of realising lateral or upward job-to-job transitions in particular, as well as their risk of becoming unemployed, rises as the regional human capital endowment grows. Employees who have received a polytechnic degree are less stably employed in regions with a higher endowment of human capital. However, their opportunities to transition into full-time or part-time employment are increased. In regions with a higher share of skilled workers, employees with a university degree are more stably employed and have lower unemployment risks. These findings suggest, on the one hand, that hypothesis 9, namely that a high local level of human capital results in less stable jobs for lower-skilled employees, must be rejected, while on the other hand, hypothesis 10, stating that the highly skilled will profit through stable jobs or even upward inter-firm mobility, holds true.

that the results do not significantly change when including imputed wages; this also applies to the estimates on upward mobility or our findings for employees with a polytechnic or university degree. In summary, all robustness tests indicate that our findings are also obtained when using other measurement concepts, different definitions for job-to-job transitions or imputed wages and thus remain stable across all specifications. 6. Discussion Recent life course research accentuates that employment trajectories are governed by individual determinants. However, individuals’ actions are always embedded within a particular framework (Coleman, 1990; Esser, 1996). For this reason, we have paid particular attention to structural effects and have extended current life course research by exploring how regional determinants impact employment trajectories. We investigated two aspects that have been frequently analysed in regional economics. First, we examined urbanisation processes (Krugman, 1991) and the associated pooling of labour in certain regions (Wheeler, 2006). Second, we focused on the regional qualification structure, which is considered a key factor for regional growth by endogenous growth theory (Lucas, 1988; Romer, 1990). We could descriptively show that the distribution of population density, as well as the shares of highly qualified employees, vary greatly among German planning regions. This is a first hint that working conditions differ systematically among regions. To assess the impact of these structural factors on employment trajectories, we estimated several Cox regression models. Coefficient estimates of different specifications with and without firm- or regionspecific variables point to selection processes of groups of workers and firms into heterogeneous regions. This finding is in line with previous research (Dütsch & Struck, 2014; Mumford & Smith, 2004). To further explore selective regional effects on specific groups of workers, we additionally included cross-level interactions in the regression models. Our first main finding is that the extent of regional agglomeration and the endowment of human capital differently affect working careers: Regarding the former, we can state that although there is less job stability in urban areas than in rural regions, opportunities for direct jobto-job transitions – and even upward inter-firm mobility – exist. There is also a greater opportunity for individuals to fill full- or part-time jobs subject to social insurance contributions. On the one hand, this finding contradicts the assumption of job search theories that greater pooling of workers in agglomerated areas leads to comparatively better matches between employers and employees and to higher job stability (Wheeler, 2006); it also indicates that findings on stronger labour demand in densely populated areas – identified by empirical studies on regional economics (Blien et al., 2006; Dauth, 2013; Farhauer & Granato, 2006), which observe the regional stock of employment but not dynamics in employment histories – cannot be transferred to research on employment trajectories. On the other hand, our finding is in line with a large body of empirical research, which also emphasises that regional agglomeration causes wage increases (Combes et al., 2008; De La Roca & Puga, 2017; Duranton & Puga, 2004, chap. 48; Kelle, 2016; Melo & Graham, 2014; Wheeler, 2008). Regarding the regional human capital endowment, a high level of endowment does not cause stable employment, but it does increase the risk of becoming unemployed in the case of a job exit. Thus, our findings do not support the theoretical assumption that human capital works as an “engine of growth” through the proximity of highly qualified and low-skilled workers (Glaeser, 1999; Jovanovic & Rob, 1989) and corresponding results of empirical studies on regional economics (Poelhekke, 2013; Schlitte, 2012; Shapiro, 2006; Südekum, 2008). However, there is obviously more competitive pressure between employees in regions with a comparatively greater share of highly skilled workers. Against the backdrop of approaches based on statistical discrimination (Arrow, 1971; Phelps, 1972), we further investigated

5.5. Robustness checks To assess to what extent the results depend on the way the determinants are measured, we performed four robustness checks. First, the measurement concept for the explanatory variables at the regional level was altered. Instead of the variable “types of regions”, which is constructed on the basis of the population share in large and medium cities, the presence and size of a big city, the population density of the planning region, and the population density of the planning region without consideration of the big cities (BBSR, 2018), we used the simple indicator “population density” provided by the BBSR (see Tables 8 and 9 in the appendix). Additionally, in Tables 10 and 11, we replaced the indicator “human capital endowment” with a simpler measure of the highly qualified employees in a region based on the employment statistics of the Federal Employment Agency. Estimations based on these alternative measurement concepts for types of regions and the human capital endowment produce very similar results to our models in Tables 2–5. There are no large deviations as far as the direction of influence of the coefficients and the significances are concerned. Second, we changed the definition of an employment gap and expanded it from 90 days to 180 days. Because this redefinition may impact all transitions, we again re-estimated the duration and competing risk models. The results in Tables 12–15 do not indicate significant differences depending on the amount of the employment gap. Third, in cases of jobto-job transitions, we modified the definition of “downward mobility” from a decrease in wages of more than 5 percent to more than 10 percent, and “upward mobility” from an increase in wages of at least 10 percent to at least 20 percent. The competing risks estimates in Tables 16–19 show no particular differences in the results achieved. However, some effects become more pronounced because the definition of upward and downward mobility has been made stricter. In particular, the probability of downward job-to-job mobility significantly decreases for employees with a university degree in urban regions (see Table 17). Fourth, because income above the income ceiling is right censored in the administrative data, we excluded these cases from our sample. This procedure could especially affect our findings on upward mobility. To test the robustness of our results, we imputed right censored wages based on a procedure by Gartner (2005).9 Tables 20–23 in turn show 9 The Tobit regression includes information about age (linear, squared), gender, education, work experience (linear, squared), firm size, and industry affiliation.

55

Advances in Life Course Research 40 (2019) 43–84

M. Dütsch, et al.

whether specific groups of employees experience disadvantages or benefits from regional characteristics. Thus, the overall positive agglomeration effect seems to be predominantly explained by females because they are more mobile between firms and able to realise lateral or upward firm-to-firm mobility in denser regions. However, they face a lower risk of filling marginal jobs. This is an important result because, according to the main effect of sex (in the estimation without interaction effects), female workers comparatively seldom manage an upward transition or a transition to a full-time job, while their probability of filling a part-time job or taking on marginal employment is increased. This main effect led us – in line with results on gendered lives (Grunow, Schulz, & Blossfeld, 2012; Treas & Lui, 2013) – to the conclusion that women often take responsibility for child-rearing and thus are disadvantaged regarding the labour market. With respect to region-specific effects, females seem to profit, to a certain degree, from living and working in denser regions. The relationship between sex and the regional human capital endowment is twofold: On the one hand, female workers obviously profit from such regions due to stable employment relations and the opportunity to shift into full-time jobs in the case of job-to-job transitions. On the other hand, they are at a greater risk of taking on marginal employment. This finding indicates that employers can be more selective in their choice of employees in regions with a high regional human capital endowment, which is in line with statistical discrimination theories (Arrow, 1971; Phelps, 1972). Foreign workers also seem to benefit from regional population density. They are more stably employed in denser regions. However, their probability of transitioning to marginal employment is higher in urban areas. This similarly applies to foreign employees in regions with a high endowment of human capital, who profit from a greater share of skilled employees in a region due to increased opportunities of upward inter-firm mobility and decreased unemployment risks. This is in contrast to the main effect of nationality in our estimation without interaction effects, which pointed – in the case of a job exit – to fewer opportunities to directly move to another employer, as well as to higher unemployment risks, for foreign employees compared to Germans. Accordingly, this type of discrimination on the labour market, which has also been found by previous research (Kogan, 2004, 2011), can apparently be diminished by these two structural factors. Furthermore, theories of discrimination emphasise that employers ascribe a depreciation of previously accumulated human capital to those employees who have experienced unemployment. Thus, these employees are disadvantaged in the labour market (Arulampalam, 2001; Manzoni & Mooi-Reci, 2011). Our findings actually show that the risk of downward mobility or transitions into unemployment increase in denser regions. Accordingly, the detrimental main effect in our estimation without interaction effects that the risks of becoming unemployed increase with the share of previous unemployment periods (see also findings in Blossfeld, 1985; Manzoni, 2012) is amplified by greater pooling of workers in specific regions. Employers seem to be more selective by negatively assessing the signalling effects of unemployment (Spence, 1973, 2002). This also explains the reduced upward mobility and fewer transitions into full-time employment in regions with a high endowment of human capital. However, in such regions, the risk of becoming unemployed is decreased, so that the picture in regions with a high local level of human capital is mixed. Our last main finding concerns the highest achieved education level. The results indicate that employees who have acquired an occupational qualification benefit more from regional agglomeration than low or even highly qualified workers: Employees who completed, at most, secondary school and vocational training benefit from working in urban areas by being more stably employed. Employees with an advanced secondary school and vocational training certificate are less often faced with lateral or downward, but more often with upward, inter-firm mobility in urban areas. Thus, the positive main effect from our estimation without interaction effects is strengthened in denser regions. Employees who completed a polytechnic degree profit most from working in an urban area due to the greater

opportunity for upward mobility, a lower risk of lateral or downward job-tojob mobility, and lower risks of transitions into marginal employment or even unemployment in the case of a job exit. This can be explained by the fact that in Germany, vocational training is highly regulated. Thus, the professional certificate has a signal function for employers (Konietzka, 2010). Obviously, this signal gains importance in the matching process in regions with a great pooling of potential employees. By contrast, according to theoretical approaches on skill segregation (Acemoglu, 1999; Duranton, 2004), employees who have completed vocational training are employed less stably and face greater risks of entering unemployment as the regional human capital endowment increases. Only employees with a university degree are both more stably employed and have lower unemployment risks. These findings suggest that in such regions, workers are, to an increasing degree, employed dependent on their qualifications in the production system, which leads to a rise in productivity of the highly skilled workforce, while the productivity of the less-skilled decreases (Acemoglu, 1999; Duranton, 2004). 7. Conclusion The employment period is considered to be a central stage in the life course (Kohli, 1985; Mayer, 2009). We have noted the fact that so far, life course research has mainly focused on analysing the effects of individual determinants – e.g., sex, age, and education – on labour market outcomes. Against the backdrop of Coleman's (1990) or Esser's (1996) claim that the broader social context is also highly relevant for understanding individuals’ actions, our objective was to enhance understanding of the effects of structural factors on employment trajectories. By focussing in greater detail on structural effects, this article extends current research on individual determinants of employment trajectories in two ways. First, to gain a fuller picture of structural determinants, we observed factors at the individual level, the firm level and the regional level. To obtain insight into this complex multi-level framework, we used a German linked employer–employee dataset (LIAB) that was merged with data on regional characteristics for 96 planning regions taken from the Federal Institute for Research on Building, Urban Affairs and Spatial Development (BBSR) and the Establishment History Panel (BHP). Second, in addition to common labour market theories, we referred to approaches in regional economics (Fujita, Krugman, & Venables, 2001; Krugman, 1991, 1998) in order to explain structural effects and to derive hypotheses. Descriptive statistics indicate that transition patterns vary between types of region. Unemployment risks in particular are lower in urban areas. Furthermore, a high human capital endowment is associated with more upward or lateral job-to-job mobility but also with higher unemployment risks. In line with previous research, multivariate analyses show that determinants on the individual as well as the firm level are significant regarding chances and risks in the life course. Additional, but in quantitative terms smaller, effects arise from the regional level. However, there are considerable regional heterogeneities regarding population density and the amount of human capital endowment, which influence working careers differently. In general, regional agglomeration predominantly offers opportunities regarding employment trajectories, while regional human capital accumulation increases employment risks. Particularly, our findings indicate that group-specific inequalities regarding employment careers can be weakened or even strengthened by regional frame conditions. With respect to sex, nationality, previous unemployment experiences and level of education, female and foreign employees benefit most from denser regions as well as from a higher endowment of human capital. By contrast, the disadvantages of employees who experienced unemployment periods during their working life are increased by both of these regional characteristics. In turn, findings regarding education level are mixed: Workers with vocational qualifications profit from regional agglomeration to a greater extent than low or even highly qualified workers. However, a high endowment of local human capital leads to skill segregation between vocationally trained and highly qualified employees. 56

Advances in Life Course Research 40 (2019) 43–84

M. Dütsch, et al.

Appendix A

Table 6 Descriptive statistics. Characteristics

Mean/sharea

Standard deviation

Sex Female Male

44.66 55.34

–/– –/–

Age (in years) Age (in years squared)

35.89 1438.358

12.23 957.22

Nationality Foreign German

6.83 93.17

–/– –/–

14.51

–/–

57.81

–/–

7.37

–/–

6.35

–/–

4.24 9.72

–/– –/–

Employment state Full-time Part-time Marginal employment

64.27 18.46 17.26

–/– –/– –/–

Daily wage (deflated)

3.71

1.04

Previous employment trajectory Previous vocational education First employment Share of full-time employment Share of part-time employment Share of unemployment Share of nonemployment Left-censored

0.01 0.07 0.42 0.14 0.18 0.17 0.01

0.1 0.25 0.38 0.25 0.27 0.27 0.08

Employment prospects for the forthcoming year The level will be approximately constant The level of employment will increase The level of employment will decrease Not sure at present

50.39 17.26 20.07 12.28

–/– –/– –/– –/–

Highest education level At most secondary school and no vocational training At most secondary school and vocational training Advanced secondary school and no vocational training Advanced secondary school and vocational training Polytechnic degree University degree

Expected development of business volume in the current year compared to previous year It is expected to remain constant 43.67 It is expected to increase 28.56 It is expected to decrease 18.09 Do not know at present 9.69

–/– –/– –/– –/–

Apprenticeships for vocational training are offered Yes No

84.67 15.27

–/– –/–

0.74 81.84

0.29 32.02

10.78

–/–

63.63 13.99 11.60

–/– –/– –/–

Works council Yes No

69.38 30.62

–/– –/–

Firm size Small firm Small-medium-sized firm

4.55 26.55

–/– –/–

Share of high-qualified employees Average gross daily wage of full-time employees Collective agreement Establishment not bound by a collective agreement Sectoral collective agreement Company collective agreement Orientation of sectoral collective agreement

(continued on next page) 57

Advances in Life Course Research 40 (2019) 43–84

M. Dütsch, et al.

Table 6 (continued) Characteristics

Mean/sharea

Standard deviation

Medium-sized firm Larger firm

21.77 47.12

–/– –/–

Sector Manufacturing industry Agriculture, forestry, and mining Construction Trade Services for firms Other services

25.00 3.83 3.41 4.71 15.13 47.92

–/– –/– –/– –/– –/– –/–

Types of region Rural area Area in urbanisation process Urban area

29.23 32.21 38.56

–/– –/– –/–

Human capital endowment Share of low-skilled workers Share of skilled workers Share of highly skilled workers Share of workers whose skill level is unknown

12.20 58.66 8.40 20.74

3.92 6.92 2.69 6.18

Unemployment rate

11.08

Number of observations

4.67 1138538

Analyses are based on all employment spells from persons who took up employment between 1.1.2000 and 31.12.2010. The period of observation ranged between 1.1.2000 and 31.12.2010. Source: LIAB (LM 9310), own calculations. a Percentages do not add up to exactly 100 due to rounding error.

Table 7 Estimates for exit from job (focus on types of regions) using different specifications. Independent variables

Region-specific factors Type of regions (Ref.: rural area) Area in urbanisation process Urban area Unemployment rate

Model 1: Exit from job (individual factors)

Model 2: Exit from job (individual and firm-specific factors)

Model 3: Exit from job (full model)

Model 4: Exit from job (stratified full model)

–/– –/– –/–

–/– –/– –/–

1.020 1.145* 0.995

–/– –/– 1.014*

0.914*** 0.975*** 1.000*** 1.206***

0.914*** 0.975*** 1.000*** 1.196***

0.925*** 0.975*** 1.000*** 1.172***

0.829*** 1.327*** 0.866*** 0.768*** 0.905**

0.834*** 1.328*** 0.868*** 0.773*** 0.911**

0.837*** 1.352*** 0.873*** 0.774*** 0.925*

0.921** 0.753*** 0.728***

0.923** 0.755*** 0.727***

0.926** 0.761*** 0.727***

0.937 1.362*** 1.670*** 1.821*** 1.569*** 1.031

0.939 1.363*** 1.667*** 1.831*** 1.566*** 1.035

0.943 1.361*** 1.661*** 1.811*** 1.554*** 1.023

1.014 1.072** 1.062*

1.005 1.068** 1.089***

0.974 1.057**

0.976 1.059**

Individual factors Sex (1 = female) 0.920*** Age (in years) 0.983*** Age (in years squared) 1.000* Nationality (1 = foreign) 1.181*** Highest education level (Ref.: at most secondary school and no vocational training) At most secondary school and vocational training 0.815*** Advanced secondary school and no vocational training 1.074 Advanced secondary school and vocational training 0.733*** Polytechnic degree 0.675*** University degree 0.777*** Employment state (Ref.: full-time) Part-time 0.888*** Marginal employment 0.668*** Daily wage (deflated) 0.661*** Previous employment trajectory (Ref.: previous vocational education) First employment 0.897 Share of full-time employment 1.233** Share of part-time employment 1.618*** Share of unemployment 2.215*** Share of nonemployment 1.521*** Left-censored 1.019

Firm-specific factors Employment prospects for the forthcoming year (Ref: will be approximately constant) The level of employment will increase –/– 1.016 The level of employment will decrease –/– 1.072** Not sure at present –/– 1.061** Expected development of business volume in the current year compared to previous year (Ref.: It is expected to remain constant) It is expected to increase –/– 0.975 It is expected to decrease –/– 1.057**

(continued on next page) 58

Advances in Life Course Research 40 (2019) 43–84

M. Dütsch, et al.

Table 7 (continued) Independent variables

Model 1: Exit from job (individual factors)

Do not know at present –/– Apprenticeships for vocational training are offered (1 = yes) –/– Works council (1 = yes) –/– Share of high-qualified employees –/– Average gross daily wage of full-time employees –/– Collective agreement (Ref.: establishment not bound by a collective agreement) Sectoral collective agreement –/– Company collective agreement –/– Orientation of sectoral collective agreement –/– Firm size (Ref.: Small firm) Small-medium-sized firm –/– Medium-sized firm –/– Larger firm –/– Sector (Ref.: Manufacturing industry) Agriculture, forestry, and mining –/– Construction –/– Trade –/– Services for firms –/– Other services –/–

Model 2: Exit from job (individual and firm-specific factors)

Model 3: Exit from job (full model)

Model 4: Exit from job (stratified full model)

0.985 0.926** 0.948* 0.903** 0.994***

0.986 0.926** 0.913* 0.950* 0.993***

1.042 0.938** 0.963 0.914** 0.993***

0.996 0.978 0.955*

0.998 0.981 0.958*

0.998 0.960 0.963

1.072** 1.029 0.976

1.075*** 1.034 0.977

1.076*** 1.042 0.969

1.299*** 1.730*** 1.202*** 1.407*** 1.165***

1.302*** 1.731*** 1.191*** 1.402*** 1.162***

1.302*** 1.696*** 1.161*** 1.390*** 1.132***

Year (Ref.: 2000) 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010

1.066*** 1.143*** 1.182*** 1.236*** 1.270*** 1.256*** 1.238*** 1.185*** 1.007 0.784***

1.072*** 1.142*** 1.172*** 1.224*** 1.277*** 1.259*** 1.271*** 1.205*** 1.037 0.859***

1.073*** 1.144*** 1.175*** 1.228*** 1.282*** 1.262*** 1.271*** 1.204*** 1.036 0.861***

1.076*** 1.139*** 1.161*** 1.221*** 1.254*** 1.250*** 1.294*** 1.257*** 1.073* 0.888**

Number of observations −2*LL (starting values) −2*LL (final estimates) AIC BIC

1138538 12018907 11897636 3070.572 3070.572

1138538 12018907 11862857 2599.020 2599.020

1138538 12018907 11862703 2599.020 2599.020

1138538 8194791 8061653 1765.579 1765.579

Analyses are based on all employment periods of persons who took up employment between 1.1.2000 and 31.12.2010. The period of observation was from 1.1.2000 to 31.12.2010. *p < 0.05; **p < 0.01; ***p < 0.001. Hazard ratios for the Cox partial-likelihood estimates; Calculation of cluster-robust standard errors for 96 planning regions. Source: LIAB (LM 9310), own calculations.

Table 8 Estimates for individual, firm-specific and region-specific factors influencing employment trajectories – Robustness check using population density. Independent variables

Exit from job

Lateral mobility

Downward mobility

Upward mobility

Transition to fulltime employment

Transition to parttime employment

Transition to marginal employment

Unemployment

Region-specific factors Population density Unemployment rate

1.000 0.998

1.010** 0.988

1.000 0.988*

1.021** 0.975***

1.013** 0.988*

1.020*** 0.979***

1.000 0.959***

1.000 1.017***

0.546*** 1.087*** 0.998*** 1.165***

2.232*** 1.117*** 0.998*** 1.264***

1.070* 0.974** 1.000** 1.351***

0.789*** 1.010* 1.000*** 1.103*

1.048

0.906**

0.746***

0.934*

0.665***

0.579***

1.671***

0.341***

1.025

0.756***

0.638***

0.710***

1.609* 2.013***

0.854 1.117

0.393*** 0.313***

0.657*** 0.790***

1.523*** 1.100

–/– 1.229*

1.024 –/–

1.027 0.412***

Individual factors 1.032 0.994 0.876** Sex (1 = female) 0.914*** Age (in years) 0.975*** 1.031*** 0.998 1.019** Age (in years squared) 1.000*** 0.999*** 1.000*** 0.999*** Nationality (1 = foreign) 1.200*** 0.873** 0.987 0.932 Highest education level (Ref.: at most secondary school and no vocational training) 0.922 0. 740*** 1.024 At most secondary school and 0.833*** vocational training *** ** *** Advanced secondary school 1.330 0.668 0.686 0.802 and no vocational training Advanced secondary school 0.868*** 0.804*** 0.596*** 1.194** and vocational training Polytechnic degree 0.774*** 0.850** 0.460*** 1.413*** University degree 0.912** 0.846* 0.464*** 1.425*** Employment state (Ref.: full-time) Part-time 0.923** 0.587*** 0.724*** 0.400*** Marginal employment 0.755*** 1.454*** 4.523*** 0.348***

(continued on next page)

59

Advances in Life Course Research 40 (2019) 43–84

M. Dütsch, et al.

Table 8 (continued) Independent variables

Exit from job

Lateral mobility

Downward mobility

Daily wage (deflated) 0.727*** 1.466*** 1.702*** Previous employment trajectory (Ref.: previous vocational education) First employment 0.939 0.833 0.734 Share of full-time employment 1.365*** 1.257 1.339 *** Share of part-time 1.671 1.652 2.682*** employment Share of unemployment 1.833*** 0.960 1.353 Share of nonemployment 1.569*** 1.094 1.400 Left-censored 1.034 0.774 0.741

Upward mobility

Transition to fulltime employment

Transition to parttime employment

Transition to marginal employment

Unemployment

0.619***

0.802***

0.719***

0.601***

0.769***

0.633** 1.043 1.521*

0.475*** 2.005*** 1.278

1.515 2.121 5.911***

1.080 2.240 9.326***

1.128 1.951*** 2.061***

0.978 1.142 0.740

0.953 0.878 0.546*

2.602 3.116* 1.354

2.837* 2.820* 0.873

4.362*** 1.333 0.821

1.090*

1.120**

1.032

0.909**

1.065

1.115*

1.105**

1.060*

1.115*

1.027

0.995 1.075 0.939 0.918

0.947 1.064* 0.982 0.954

Firm-specific factors Employment prospects for the forthcoming year (Ref: will be approximately constant) The level of employment will 1.014 1.057 1.134** 1.160** increase The level of employment will 1.072** 1.216*** 1.196*** 1.060 decrease Not sure at present 1.062* 1.113 0.994 1.029 Expected development of business volume in the current year compared to previous year (Ref.: * It is expected to increase 0.975 1.103 1.021 1.098* 1.120* 1.091* 1.076 It is expected to decrease 1.057** Do not know at present 0.983 1.346*** 0.950 0.947 1.010 1.095 0.981 Apprenticeships for vocational 0.927** training are offered (1 = yes) Share of high-qualified 0.913* 0.930 0.689*** 0.942 employees * * Works council (1 = yes) 0.949 0.914 1.017 0.953 0.990*** 0.990*** 0.990*** Average gross daily wage of full- 0.993*** time employees Collective agreement (Ref.: establishment not bound by a collective agreement) Sectoral collective agreement 0.999 1.018 1.079 1.098 Company collective 0.978 0.951 0.885 0.980 agreement Orientation of sectoral 0.957* 0.942 0.990 0.930 collective agreement Firm size (Ref.: small firm) 0.941 1.055 1.118 Small-medium-sized firm 1.074** Medium-sized firm 1.034 0.777** 0.946 0.999 Larger firm 0.977 0.720** 0.827 0.920 Sector (Ref.: manufacturing industry) *** Agriculture, forestry, and 1.306 1.427 1.242 1.049 mining Construction 1.730*** 1.706*** 1.404*** 1.103 Trade 1.195*** 1.735*** 1.114 1.412*** Services for firms 1.404*** 2.314*** 1.588*** 2.559*** Other services 1.163*** 1.749*** 1.088 1.490***

0.978 0.996 it is expected to remain constant) 1.034 0.954 1.018 1.067 * 0.841 0.901 1.022 1.088* 0.676***

0.650***

0.650*** *

0.977

0.997 0.991***

1.104 0.990***

0.904 0.990***

0.928 0.987***

0.934 0.949

0.996 0.896

1.005 1.076

1.020 0.933

0.957

0.962

0.924

0.986

1.020 0.937 0.861*

1.057 0.894 0.831*

1.054 1.049 1.040

0.931 0.834** 0.731***

1.192

1.496*

1.330*

1.159

1.172 1.335** 1.671*** 1.209*

1.300 1.938*** 2.517*** 2.237***

1.241* 1.142 1.410*** 1.023

2.222*** 1.049 1.300*** 1.003

Year (Ref.: 2000) 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010

1.072*** 1.144*** 1.177*** 1.229*** 1.285*** 1.263*** 1.270*** 1.200*** 1.033 0.857***

0.988 0.933 1.016 1.074 1.072 1.044 0.983 0.875* 0.661*** 0.410***

0.935 0.919 0.941 1.015 1.146* 1.188* 1.122 0.988 0.827** 0.541***

0.988 0.948 0.939 1.106 1.456*** 1.517*** 1.463*** 1.168* 1.070 0.762**

0.944 0.987 1.136*** 1.222*** 1.548*** 1.698*** 1.563*** 1.365*** 1.288*** 1.031

1.003 1.038 1.003 1.081 1.401*** 1.410*** 1.367*** 1.364*** 1.225*** 0.912

1.086 1.397*** 2.074*** 2.398*** 2.857*** 2.944*** 2.842*** 2.595*** 2.199*** 1.486***

1.146*** 1.299*** 1.344*** 1.154*** 0.856** 0.787*** 0.886 0.877 0.789** 1.045

Number of observations −2*LL (starting values) −2*LL (final estimates) AIC BIC

1138538 12018907 11862773 2599.021 2599.021

1138538 577244 570496 1141102 1141750

1138538 599318 587881 1175871 1176519

1138538 1015802 989049 1978206 1978855

1138538 420843 369677 739461 740097

1138538 255618 238156 476420 477056

1138538 405090 381732 763571 764207

1138538 3707672 3579991 7160091 7160739

Analyses are based on all employment periods of persons who took up employment between 1.1.2000 and 31.12.2010. The period of observation was from 1.1.2000 to 31.12.2010. *p < 0.05; **p < 0.01; ***p < 0.001. Hazard ratios for the Cox partial-likelihood estimates; Calculation of cluster-robust standard errors for 96 planning regions. Source: LIAB (LM 9310), own calculations.

60

1.000*** 1.382 1.210**

1.000*** 1.004 0.998

1.000** 0.994 1.000

1.000*** 1.225*** 1.000

1.000*** 0.543** 1.000

61

0.988*

0.987*

0.991* 1.000 0.989*

0.854** 1.000 0.975**

***

0.996 0.998

0.993*

0.987

**

0.072*

0.775 2.116*** 0.608*** 0.400*** 0.320***

1.000

1.000 2.887* 1.000

1.000 1.235*** 1.012***

1.000 1.137*** 0.901***

Transition to marginal employment

0.879* 0.999

1.000

1.000

1.000

0.950 0.332*** 0.679*** 0.636*** 0.782***

1.000

1.000 4.249*** 1.190**

1.000*** 1.096 1.000

1.000*** 0.773*** 1.000

Unemployment

Analyses are based on all employment periods of persons who took up employment between 1.1.2000 and 31.12.2010. The period of observation was from 1.1.2000 to 31.12.2010. Models additionally contain all explanatory variables that are reported in Table 2. *p < 0.05; **p < 0.01; ***p < 0.001. Hazard ratios for the Cox partial-likelihood estimates; Calculation of cluster-robust standard errors for 96 planning regions. Source: LIAB (LM 9310), own calculations.

0.994

0.997

*

0.924 0.534*** 0.716*** 0.908 1.264**

1.000***

1.000*** 2.655 1.000

1.000*** 1.250*** 1.000

1.000*** 2.345*** 1.000

Transition to part-time employment

1.083 0.650*** 0.987 1.169* 2.210***

1.000***

1.000 1.852*** 1.000

Population density Share of unemployment periods Interaction term: share of unemployment periods × population density

1.000*** 0.991 1.000

1.000*** 0.951 1.000

1.000*** 0.818*** 1.021*

Transition to full-time employment

Population density 1.000 1.000 1.000* 1.000 Highest education level (Ref.: at most secondary school and no vocational training) *** ** *** At most secondary school and vocational training 0.833 0.859 0.732 0.962 Advanced secondary school and no vocational training 1.459*** 0.693* 0.735* 0.849 Advanced secondary school and vocational training 0.848*** 0.747*** 0.585*** 1.070 Polytechnic degree 0.732*** 0.768*** 0.449*** 1.264** University degree 0.990 0.834* 0.491*** 1.432*** Interaction term: Highest education level (Ref.: at most secondary school and no vocational training) × population density Interaction term: At most secondary school and vocational 1.000 0.950** 1.000 1.000* training × population density Interaction term: Advanced secondary school and no vocational 1.000 1.001 0.998 0.995* training × population density Interaction term: Advanced secondary school and vocational 1.000 0.988* 0.974** 1.000 training × population density Interaction term: Polytechnic degree × population density 1.000 0.987* 0.971** 1.010** Interaction term: University degree × population density 0.998 1.000 1.000* 1.000

1.000 1.260*** 0.970*

Population density Nationality (1 = foreign) Interaction term: Nationality (1 = foreign) × population density

1.000*** 1.008 1.000

1.000*** 0.989 1.051*

Upward mobility

1.000*** 0.918 1.000

0.999 0.891*** 1.090**

Population density Sex (1 = female) Interaction term: sex (1 = female) × population density

Downward mobility

Lateral mobility

1.000*** 0.993 1.000

Exit from job

Independent variables

Table 9 Estimates for cross-level interactions between population density and individual factors influencing employment trajectories.

M. Dütsch, et al.

Advances in Life Course Research 40 (2019) 43–84

1.007 0.997

Region-specific factors Human capital endowment (share of highly skilled workers) Unemployment rate

62

0.992 0.998 1.000*** 0.990 0.725*** 0.678*** 0.587*** 0.443*** 0.448*** 0.722*** 4.603*** 1.729*** 0.721 1.314 2.652*** 1.314 1.392 0.732

0.907* 0.666** 0.794*** 0.823*** 0.821** 0.588*** 1.483*** 1.486*** 0.822 1.235 1.636 0.939 1.090 0.769

1.017 0.991

Downward mobility

1.030 1.032*** 1.000*** 0.871**

1.020 0.999

Lateral mobility

Firm-specific factors Employment prospects for the forthcoming year (Ref: will be approximately constant) The level of employment will increase 1.014 1.057 1.138*** The level of employment will decrease 1.071** 1.220*** 1.205*** Not sure at present 1.061** 1.104 0.984 Expected development of business volume in the current year compared to previous year (Ref.: it is expected to remain constant) It is expected to increase 0.974 1.106* 1.022 It is expected to decrease 1.058** 1.132* 1.105* Do not know at present 0.987 1.359*** 0.959 Apprenticeships for vocational training are offered (1 = yes) 0.927** 1.009 1.083 * Share of high-qualified employees 0.907 0.896 0.659*** 1.008 Works council (1 = yes) 0.950 0.907* Average gross daily wage of full-time employees 0.993*** 0.991*** 0.990*** Collective agreement (Ref.: establishment not bound by a collective agreement) Sectoral collective agreement 1.003 1.018 1.067 Company collective agreement 0.983 0.977 0.911 * 0.946 0.983 Orientation of sectoral collective agreement 0.960

Individual factors Sex (1 = female) 0.914*** Age (in years) 0.976*** Age (in years squared) 1.000*** Nationality (1 = foreign) 1.196*** Highest education level (Ref.: at most secondary school and no vocational training) At most secondary school and vocational training 0.832*** Advanced secondary school and no vocational training 1.330*** Advanced secondary school and vocational training 0.867*** Polytechnic degree 0.769*** University degree 0.908** Employment state (Ref.: full-time) Part-time 0.923** Marginal employment 0.757*** Daily wage (deflated) 0.728*** Previous employment trajectory (Ref.: previous vocational education) First employment 0.936 Share of full-time employment 1.357*** Share of part-time employment 1.665*** Share of unemployment 1.822*** Share of nonemployment 1.564*** Left-censored 1.031

Exit from job

Independent variables

1.092* 1.068 0.965 1.038 1.024 0.859 1.012 0.649*** 1.005 0.992*** 0.935 0.979 0.969

1.098* 1.087* 0.961 0.977 0.898 0.945 0.990*** 1.107 1.018 0.938

0.472** 1.957** 1.267 0.919 0.876 0.541*

0.624** 1.022 1.502* 0.951 1.134 0.734

1.160** 1.062 1.019

3.967*** 1.102 0.805***

1.029 0.652*** 1.005 1.568* 1.978***

0.547*** 1.090*** 0.998*** 1.151***

1.033*** 0.987*

Transition to fulltime employment

0.401*** 0.347*** 0.620***

1.010 0.811 1.187** 1.364*** 1.392***

0.872** 1.021*** 0.999*** 0.926

1.027* 0.975***

Upward mobility

0.998 0.928 0.968

0.955 1.082* 0.911 1.072 0.621*** 1.100 0.991***

1.123** 1.124* 0.987

1.477 2.055 5.774*** 2.494 3.082* 1.325

–/– 1.234* 0.725***

0.886*** 0.577*** 0.747*** 0.810* 1.079

2.221*** 1.118*** 0.998*** 1.261***

1.028** 0.980***

Transition to parttime employment

Table 10 Estimates for individual, firm-specific and region-specific factors influencing employment trajectories – Robustness check using the share of highly skilled workers.

1.018 1.075 0.931

0.995 1.073 0.943 0.927 0.653*** 0.909* 0.990***

1.029 1.100** 1.121*

1.075 2.227 9.265*** 2.858* 2.801* 0.876

1.033 –/– 0.602***

0.753*** 1.633*** 0.639*** 0.394*** 0.315***

1.072* 0.975*** 1.001** 1.330***

1.007 0.959***

Transition to marginal employment

(continued on next page)

1.036 0.938 0.996

0.947 1.062* 0.986 0.963 0.979 0.933 0.987***

0.905** 1.058* 1.028

1.121 1.933*** 2.044*** 4.341*** 1.318 0.820

1.033 0.414*** 0.767***

0.942 0.339*** 0.711*** 0.660*** 0.793***

0.789*** 1.010* 1.000*** 1.086*

1.019* 1.015***

Unemployment

M. Dütsch, et al.

Advances in Life Course Research 40 (2019) 43–84

1.420 1.716*** 1.750*** 2.329*** 1.765***

1.305*** 1.727*** 1.191*** 1.400*** 1.161*** 1.071*** 1.142*** 1.173*** 1.225*** 1.279*** 1.256*** 1.259*** 1.190*** 1.021 0.847*** 1138538 12018907 11862675 2599 2599

Year (Ref.: 2000) 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010

Number of observations −2*LL (starting values) −2*LL (final estimates) AIC BIC

63

1138538 599318 588091 1176291 1176940

0.929 0.904 0.917 0.990 1.106 1.157 1.086 0.964 0.795** 0.522***

1.226 1.423*** 1.138 1.611*** 1.109

1.047 0.925 0.824

Downward mobility

1138538 1015802 989191 1978490 1979138

0.985 0.935 0.919 1.085 1.415*** 1.476*** 1.409*** 1.130 1.016 0.728***

1.036 1.105 1.420*** 2.554*** 1.498***

1.110 0.978 0.910

Upward mobility

1138538 420843 369682 739471 740108

0.941* 0.977 1.118** 1.201*** 1.506*** 1.644*** 1.503*** 1.318*** 1.221** 0.979

1.166 1.157 1.331** 1.673*** 1.206*

1.001 0.911 0.848**

Transition to fulltime employment

1138538 255618 238225 476557 477193

0.996 1.023 0.979 1.056 1.354*** 1.373*** 1.320*** 1.326*** 1.165* 0.872

1.470* 1.308 1.959*** 2.531*** 2.265***

1.041 0.863 0.813*

Transition to parttime employment

1138538 405090 381750 763607 764243

1.087 1.399*** 2.074*** 2.395*** 2.863*** 2.927*** 2.818*** 2.566*** 2.172*** 1.470***

1.336* 1.225* 1.112 1.390*** 1.000

1.046 1.048 1.028

Transition to marginal employment

1138538 3707672 3580072 7160254 7160902

1.145*** 1.297*** 1.349*** 1.149*** 0.854** 0.781*** 0.878 0.866 0.778** 1.032

1.161 2.198*** 1.032 1.288*** 0.987

0.927 0.836** 0.725***

Unemployment

Analyses are based on all employment periods of persons who took up employment between 1.1.2000 and 31.12.2010. The period of observation was from 1.1.2000 to 31.12.2010. *p < 0.05; **p < 0.01; ***p < 0.001. Hazard ratios for the Cox partial-likelihood estimates; Calculation of cluster-robust standard errors for 96 planning regions. Source: LIAB (LM 9310), own calculations.

1138538 577244 570590 1141288 1141936

0.983 0.923 0.994 1.054 1.043 1.019 0.948 0.850* 0.633*** 0.396***

0.936 0.763** 0.718**

Lateral mobility

1.072** 1.031 0.974

Exit from job

Firm size (Ref.: small firm) Small-medium-sized firm Medium-sized firm Larger firm Sector (Ref.: manufacturing industry) Agriculture, forestry, and mining Construction Trade Services for firms Other services

Independent variables

Table 10 (continued)

M. Dütsch, et al.

Advances in Life Course Research 40 (2019) 43–84

1.018 1.398 0.993

1.024* 1.237 0.991

1.006 1.698*** 1.007

Human capital endowment (share of highly skilled workers) Share of unemployment periods Share of unemployment periods × human capital endowment (share of highly skilled workers)

64

1.006

1.077 1.053* 1.005 1.006

1.009 1.064*** 1.029 0.986

0.984

1.000

1.000

1.000

1.052

0.942*

0.996

0.977

1.051** 0.980

0.987

1.013

0.901**

1.015

1.037*

1.031**

1.020* 0.987 1.040

0.750* 0.208*** 0.502*** 0.578* 0.990

1.014

1.009 4.42*** 0.969*

1.019* 1.047 0.998

1.005 0.725*** 1.009

Unemployment

0.772* 3.061*** 0.727 0.499*** 0.329***

1.031

1.038**

1.005 2.514* 1.014

1.003 0.968 1.002

1.021* 1.383** 0.994

Transition to marginal employment

1.002 0.395*** 0.534*** 0.462** 1.314

1.028** 2.467 1.000

1.034*** 0.969 0.980*

1.026 1.030 1.020*

1.050*** 3.016*** 0.996

1.031** 0.529*** 1.024* 1.034*** 1.246 0.992

Transition to part-time employment

Transition to full-time employment

0.662*** 0.823 1.059 0.357** 0.387* 0.671* 0.409 0.714 0.995 0.628 1.328 0.696 0.958 1.324 3.141*** capital endowment (share of highly skilled workers) 1.012 1.012 1.000

1.000

1.030* 0.901 0.975*

1.028* 1.113 0.982

1.022 0.778* 1.012

Upward mobility

Analyses are based on all employment periods of persons who took up employment between 1.1.2000 and 31.12.2010. The period of observation was from 1.1.2000 to 31.12.2010. Models additionally contain all explanatory variables that are reported in Table 3. *p < 0.05; **p < 0.01; ***p < 0.001. Hazard ratios for the Cox partial-likelihood estimates; Calculation of cluster-robust standard errors for 96 planning regions. Source: LIAB (LM 9310), own calculations.

Human capital endowment (share of highly skilled workers) 1.003 1.000 Highest education level (Ref.: at most secondary school and no vocational training) At most secondary school and vocational training 0.781*** 0.741 Advanced secondary school and no vocational training 1.369 0.276*** *** Advanced secondary school and vocational training 0.719 0.428*** Polytechnic degree 0.705* 0.628 University degree 1.047 0.958 Interaction term: Highest education level (Ref.: at most secondary school and no vocational training) × human 1.012 Interaction term: At most secondary school and vocational 1.017* training × human capital endowment (share of highly skilled workers) Interaction term: Advanced secondary school and no vocational 0.997 1.005 training × human capital endowment (share of highly skilled workers) Interaction term: Advanced secondary school and vocational 1.019* 1.064*** training × human capital endowment (share of highly skilled workers) Interaction term: Polytechnic degree × human capital endowment 1.018* 1.029 (share of highly skilled workers) * Interaction term: University degree × human capital endowment (share 0.976 0.986 of highly skilled workers)

1.018 1.098 0.989

1.022 1.059 0.980

1.009* 1.492* 0.978

Human capital endowment (share of highly skilled workers) Nationality (1 = foreign) Nationality (1 = foreign) × human capital endowment (share of highly skilled workers)

1.017 0.978 1.001

1.016 0.941 1.009

1.003 1.241*** 0.990*

Downward mobility

Human capital endowment (share of highly skilled workers) Sex (1 = female) Sex (1 = female) × human capital endowment (share of highly skilled workers)

Lateral mobility

Exit from job

Independent variables

Table 11 Estimates for cross-level interactions between the share of highly skilled workers and individual factors influencing employment trajectories.

M. Dütsch, et al.

Advances in Life Course Research 40 (2019) 43–84

1.028 1.160* 0.999

Region-specific factors Type of region (Ref.: rural area) Area in urbanisation process Urban area Unemployment rate 1.028 1.199* 0.994

Lateral mobility

65

2.480*** 1.121 0.782*** 0.448*** 1.742** 1.167

0.823 1.099 1.278*** 1.325*** 0.379*** 0.318*** 0.575*** 0.682* 1.212 1.620** 1.054 1.263 0.791

0.674*** 0.543*** 0.406*** 0.407*** 0.711*** 4.762*** 1.737*** 0.756 1.370 2.586*** 1.447 1.530 0.806

3.901*** 1.154 0.742

2.673* 2.769* 0.956

0.653***

0.688*** 1.018 0.991***

0.955 1.052 0.938 1.104**

1.039 1.035 0.867* 1.016

1.111 0.990***

*

0.095* 0.993

0.981

0.930 0.991***

0.671***

(continued on next page)

0.933 0.988***

0.959

0.943 1.066* 0.974 0.954

1.018

1.110**

1.122* 1.069

0.997 1.077 0.963 0.951

1.065*

0.985

0.890**

1.040 1.202 1.960**

1.009 0.377*** 0.742***

0.637*** 0.750***

0.681***

0.337***

0.920*

0.797*** 1.014** 1.000*** 1.086

1.015 0.909 1.016**

Unemployment

1.220 2.164 7.937***

1.018 –/– 0.565***

0.367*** 0.277***

0.626***

1.535***

0.722***

1.095*** 0.975** 1.000** 1.296***

1.082 1.084 0.959***

Transition to marginal employment

1.090*

2.331 2.884* 1.251

1.434 1.931 4.969***

–/– 1.076 0.690***

0.754* 1.018

0.714***

0.643***

0.868***

2.151*** 1.122*** 0.998*** 1.242***

1.069 1.356*** 0.988*

Transition to parttime employment

1.088*

0.884 0.838 0.548**

1.512* 1.769***

1.021

0.744***

0.995

0.966

0.689***

0.545*** 1.090*** 0.998*** 1.153***

1.158*** 1.352*** 0.998

Transition to fulltime employment

0.875*** 1.021*** 1.000*** 0.928

1.073 1.270*** 0.982**

Upward mobility

0.990 0.997 1.000*** 1.006

1.126 1.044 0.996

Downward mobility

Firm-specific factors Employment prospects for the forthcoming year (Ref: will be approximately constant) The level of employment will 0.997 1.037 1.136*** 1.132** increase The level of employment will 1.085** 1.226*** 1.201*** 1.073 decrease Not sure at present 1.076** 1.100 0.986 1.016 Expected development of business volume in the current year compared to previous year (Ref.: it is expected to remain constant) * It is expected to increase 0.983 1.095 1.020 1.097** It is expected to decrease 1.067** 1.106 1.079* 1.086* Do not know at present 0.991 1.315*** 0.928 0.955 Apprenticeships for vocational 0.927** 1.032 1.097 1.005 training are offered (1 = yes) Share of high-qualified 0.884** 0.906 0.672*** 0.901 employees * * Works council (1 = yes) 0.952 0.927 1.021 0.953 Average gross daily wage of full0.992*** 0.990*** 0.990*** 0.990*** time employees

Individual factors Sex (1 = female) 0.883*** 1.035 Age (in years) 0.979*** 1.029*** Age (in years squared) 1.000** 0.999*** Nationality (1 = foreign) 1.150*** 0.875** Highest education level (Ref.: at most secondary school and no vocational training) At most secondary school and 0.808*** 0.885* vocational training Advanced secondary school 1.276*** 0.663*** and no vocational training Advanced secondary school 0.837*** 0.745*** and vocational training Polytechnic degree 0.755*** 0.773*** University degree 0.904** 0.782*** Employment state (Ref.: full-time) 0.575*** Part-time 0.924** Marginal employment 0.754*** 1.544*** Daily wage (deflated) 0.737*** 1.483*** Previous employment trajectory (Ref.: previous vocational education) First employment 0.899 0.798 Share of full-time employment 1.268** 1.279 *** Share of part-time 1.624 1.618 employment Share of unemployment 1.735*** 0.943 Share of nonemployment 1.371*** 1.122 Left-censored 0.980 0.770

Exit from job

Independent variables

Table 12 Estimates for individual, firm-specific and region-specific factors (focus on types of regions) influencing employment trajectories with a redefined gap (more than 180 days without employment or registered unemployment).

M. Dütsch, et al.

Advances in Life Course Research 40 (2019) 43–84

1.359

1.282**

1.741 1.201*** 1.418*** 1.132***

1.072*** 1.146*** 1.183*** 1.251*** 1.292*** 1.270*** 1.284*** 1.221*** 1.047 0.876***

1084133 10623447 10478727 2302.929 2302.929

Year (Ref.: 2000) 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010

Number of observations −2*LL (starting values) −2*LL (final estimates) AIC BIC

66

***

1084133 1305563 1278126 1278236 1278890

0.940 0.937 0.937 1.030 1.189** 1.279*** 1.229*** 1.107* 0.914 0.576***

1.394 1.131 1.577*** 1.076

1.172

1084133 2162208 2104294 2104404 2105058

0.999 0.979 0.954 1.149* 1.545*** 1.628*** 1.588*** 1.301*** 1.914* 0.839

1.088 1.404*** 2.467*** 1.443***

1.032

1.139* 1.002 0.938

0.941

1.088 1.005

Upward mobility

1084133 964066 850422 850530.2 851172.6

0.956 1.003 1.147*** 1.242*** 1.619*** 1.807*** 1.665*** 1.446*** 1.404*** 1.097

1.166 1.291* 1.595*** 1.139*

1.130

1.037 0.956 0.871*

0.978

0.929 1.004

Transition to fulltime employment

1084133 574364 536660 536767.7 537410.1

1.011 1.033 1.030 1.102 1.455*** 1.504*** 1.462*** 1.525*** 1.337*** 1.031

1.285 1.825*** 2.376*** 2.095***

1.439*

1.070 0.904 0.855*

0.983

1.005 0.963

Transition to parttime employment

1084133 963832 906328 906436.4 907078.8

1.132** 1.431*** 2.006*** 2.435*** 3.122*** 3.202*** 3.069*** 2.589*** 2.395*** 1.593***

1.125 1.124 1.288** 0.925

1.220

1.065 1.059 1.053

0.920*

1.010 1.068

Transition to marginal employment

1084133 7214303 6969219 6969329 6969983

1.151*** 1.317*** 1.365*** 1.181*** 0.868* 0.813*** 0.915 0.903 0.821* 1.147*

2.147*** 1.044 1.267** 0.985

1.140

0.932 0.842** 0.748***

0.985

1.025 0.931

Unemployment

Analyses are based on all employment periods of persons who took up employment between 1.1.2000 and 31.12.2010. The period of observation was from 1.1.2000 to 31.12.2010. *p < 0.05; **p < 0.01; ***p < 0.001. Hazard ratios for the Cox partial-likelihood estimates; Calculation of cluster-robust standard errors for 96 planning regions. Source: LIAB (LM 9310), own calculations.

1084133 1163982 1150766 1150876 1151531

0.995 0.938 1.020 1.100 1.134 1.117 1.064 0.959 0.740*** 0.460***

1.672 1.676*** 2.276*** 1.707***

***

0.944 0.768** 0.726***

1.074* 1.023 0.944

***

0.969

0.953

0.955* 1.052 0.930 0.832

1.041 0.908

not bound by a collective agreement) 0.993 1.012 0.963 0.985

Downward mobility

Collective agreement (Ref.: establishment Sectoral collective agreement Company collective agreement Orientation of sectoral collective agreement Firm size (Ref.: small firm) Small-medium-sized firm Medium-sized firm Larger firm Sector (Ref.: manufacturing industry) Agriculture, forestry, and mining Construction Trade Services for firms Other services

Lateral mobility

Exit from job

Independent variables

Table 12 (continued)

M. Dütsch, et al.

Advances in Life Course Research 40 (2019) 43–84

1.014 1.231* 0.955 0.975 0.987

0.950 1.030

1.218 1.467

0.848* 0.855** 1.021 1.069 1.746***

1.014 1.194* 0.703

1.040 1.275***

1.068* 1.139*** 1.032 1.070* 1.302***

1.010 1.076 0.920

0.998 1.001 0.824***

Exit from job Lateral mobility

67 **

*

0.631*** 1.089 0.612***

1.419** 1.295** 0.728*

0.823 0.782 0.630**

0.714 0.707* 0.797** 0.765

0.718

*

0.868 0.709* 0.759

0.676 0.812 0.765** 0.988

0.742

*

0.673**

0.812*

0.837 2.106*** 0.727*** 0.480*** 0.339***

1.208 1.322**

1.109 0.849

1.059 1.109 2.714*

1.335 1.615*

1.075 1.051 0.884

0.920 0.782***

1.131* 1.232* 1.239***

Transition to marginal employment

0.897 0.454*** 1.005

0.872 1.075 0.922 0.795

0.811

0.845

1.008

0.944 0.487*** 0.752** 0.701*** 0.726**

1.020 0.966

1.151* 1.394***

0.961 0.818*** 3.412***

1.168 1.259

1.012 0.899 0.913

1.038 1.082

1.000 0.881 0.772***

Unemployment

Analyses are based on all employment periods of persons who took up employment between 1.1.2000 and 31.12.2010. The period of observation was from 1.1.2000 to 31.12.2010. Models additionally contain all explanatory variables that are reported in Table 2. *p < 0.05; **p < 0.01; ***p < 0.001. Hazard ratios for the Cox partial-likelihood estimates; Calculation of cluster-robust standard errors for 96 planning regions. Source: LIAB (LM 9310), own calculations.

0.834 0.755* 0.740*** 0.593**

0.969 0.717* 1.113 0.729

0.693

0.666**

0.520* 0.681

0.834*

0.791**

1.007

0.502***

1.026 1.017 0861 0.934 1.452**

1.037 1.113 1.137 1.678*** 2.452***

1.306** 1.553***

1.178 1.168

0.925 1.249 1.149 1.256* 1.726***

1.106 1.195

1.026 1.307*** 2.091

1.134 1.139

1.064 1.347*** 1.105

0.884 0.888

1.164 1.471*** 2.357***

Transition to part-time employment

1.433*** 1.781***

1.252 1.225

1.410** 1.336**

1.118 1.247*** 0.748

0.844 0.841

1.164*** 1.360*** 1.253

0.997 0.889

1.162** 1.438*** 1.170***

Transition to full-time employment

1.120 1.205

1.019 1.215** 0.911

1.279 1.036

1.061 1.274* 0.850

1.028 1.298***

0.860* 1.274* 0.836*

Upward mobility

1.102 1.055 1.157

1.179 1.109

1.118 1.141*** 0.999

1.082 1.116

1.087 1.085* 0.920

Downward mobility

Type of region (Ref.: rural area) Area in urbanisation process 1.126 1.217 1.286*** Urban area 1.140** 1.231* 1.367*** Highest education level (Ref.: at most secondary school and no vocational training) At most secondary school and vocational training 0.878** 0.936 0.737*** Advanced secondary school and no vocational training 1.464*** 1.365* 1.111 Advanced secondary school and vocational training 0.853** 0.878 0.619*** *** Polytechnic degree 0.781 0.825 0.528*** University degree 0.955 0.922 0.572*** Highest education level (Ref.: at most secondary school and no vocational training) × type of region (Ref.: rural area) 0.894 At most secondary school and vocational training × area 0.883 0.627** undergoing urbanisation process Advanced secondary school and no vocational training × area 0.899 0.959 0.527** undergoing urbanisation process Advanced secondary school and vocational training × area 0.894 0.890 0.807** undergoing urbanisation process Polytechnic degree × area undergoing urbanisation process 0.793* 0.789* 0.648*** University degree × area undergoing urbanisation process 1.014 0.879 0.753 At most secondary school and vocational training × urban area 0.895* 1.019 0.951 Advanced secondary school and no vocational training × urban 1.017 0.830 0.856 area ** Advanced secondary school and vocational training × urban area 0.859 0.552 0.881** Polytechnic degree × urban area 0.929 0.216*** 0.750** University degree × urban area 0.918 0.971 0.909

Type of region (Ref.: rural area) Area in urbanisation process Urban area Share of unemployment periods Share of unemployment periods × type of region (Ref.: rural area) Share of unemployment periods × area in urbanisation process Share of unemployment periods × urban area

Type of region (Ref.: rural area) Area in urbanisation process Urban area Nationality (1 = foreign) Nationality (1 = foreign) × type of region (Ref.: rural area) Foreign × area in urbanisation process Foreign × urban area

Type of region (Ref.: rural area) Area in urbanisation process Urban area Sex (1 = female) Sex (1 = female) × type of region (Ref.: rural area) Female × area in urbanisation process Female × urban area

Independent variables

Table 13 Estimates for cross-level interactions between types of regions and individual factors influencing employment trajectories with a redefined gap (more than 180 days without employment or registered unemployment).

M. Dütsch, et al.

Advances in Life Course Research 40 (2019) 43–84

Exit from job

68 0.454*** 1.753** 1.189

1.129** 1.301*** 1.359*** 0.379*** 0.318*** 0.575*** 0.688** 1.222 1.634** 1.067 1.273 0.795

0.551*** 0.415*** 0.421*** 0.711*** 4.676*** 1.717*** 0.768 1.392 2.623*** 1.474 1.547 0.813

Firm-specific factors Employment prospects for the forthcoming year (Ref: will be approximately constant) The level of employment will 0.994 1.024 1.084* 1.117* increase The level of employment will 1.083*** 1.222*** 1.173** 1.070 decrease ** Not sure at present 1.076 1.100 0.974 1.014 Expected development of business volume in the current year compared to previous year (Ref.: it is expected to remain constant) * It is expected to increase 0.983 1.096 1.023 1.097* 1.124* 1.078* 1.083* It is expected to decrease 1.068*** Do not know at present 0.992 1.301*** 0.932 0.950 1.030 1.099 1.004 Apprenticeships for vocational 0.930** training are offered (1 = yes) Share of high-qualified 0.880** 0.924 0.691*** 0.913 employees Works council (1 = yes) 0.944* 0.927* 1.016 0.951

2.530*** 1.071 0.784***

0.849

0.708***

3.840*** 1.146 0.738

2.619* 2.753* 0.939

1.095* 0.982 0.955 1.047 0.933 1.072 0.665*** 1.092

0.980 1.033 1.035 0.879 1.019 0.684*** 1.010

0.927

0.658***

(continued on next page)

0.936

0.970

0.942 1.065* 0.985 0.967

1.024

1.113***

1.108* 1.065

1.002 1.078 0.967 0.956

1.062*

0.984

0.888**

1.027 1.021 1.938***

1.008 0.382*** 0.745***

0.627*** 0.740***

0.672***

0.334***

0.913**

0.797*** 1.014** 1.000*** 1.099*

1.004

1.020* 1.022*

Unemployment

1.208 2.145 7.900***

1.014 –/– 0.565***

0.364*** 0.277***

0.623***

1.543***

0.713***

1.092** 0.975** 1.000** 1.306***

0.949***

1.016 1.017*

Transition to marginal employment

1.066

2.372 2.915* 1.268

1.453 1.952 5.039***

–/– 1.199* 0.689***

0.775 1.050

0.729***

0.685***

0.877***

2.136*** 1.121*** 0.998*** 1.251***

0.982*

0.997 1.030****

Transition to parttime employment

1.087*

0.892 0.852 0.554**

1.503* 1.800***

1.027

0.762***

0.992

0.976

0.697***

0.544*** 1.090*** 0.998*** 1.152**

0.998

0.979* 0.871*** 1.020*** 0.999*** 0.927

0.997

0.994

0.979** 1.021**

Transition to fulltime employment

0.995 1.023

Upward mobility

0.985 0.996 1.000*** 1.002

0.987 1.011

Downward mobility

0.993 1.011

Lateral mobility

Individual factors Sex (1 = female) 0.882*** 1.033 Age (in years) 0.979*** 1.028*** Age (in years squared) 1.000** 0.999*** Nationality (1 = foreign) 1.150*** 0.875* Highest education level (Ref.: at most secondary school and no vocational training) 0.896* At most secondary school and 0.808*** vocational training Advanced secondary school 1.287*** 0.683** and no vocational training Advanced secondary school 0.837*** 0.754*** and vocational training Polytechnic degree 0.753*** 0.791** University degree 0.906* 0.806** Employment state (Ref.: full-time) Part-time 0.934* 0.575*** 1.504*** Marginal employment 0.758*** Daily wage (deflated) 0.738*** 1.469*** Previous employment trajectory (Ref.: previous vocational education) First employment 0.899 0.807 Share of full-time employment 1.265*** 1.295 *** 1.621* Share of part-time 1.624 employment Share of unemployment 1.731*** 0.954 Share of nonemployment 1.371*** 1.126 Left-censored 0.979 0.774

Region-specific factors Human capital endowment (Ref.: share of low-skilled workers) Share of skilled workers 1.004 Share of highly skilled 1.010 workers Unemployment rate 0.995

Independent variables

Table 14 Estimates for the individual, firm-specific and region-specific factors (focus on human capital endowment) influencing employment trajectories with a redefined gap (more than 180 days without employment or registered unemployment).

M. Dütsch, et al.

Advances in Life Course Research 40 (2019) 43–84

0.940 0.779** 0.719*** 1.383 1.632*** 1.653*** 2.203*** 1.660***

1.071** 1.022 0.938

1.288***

1.730*** 1.195*** 1.405*** 1.127***

1.067*** 1.138*** 1.173*** 1.235*** 1.271*** 1.238*** 1.235*** 1.164*** 0.990 0.824***

1084133 21246895 20957107 2302.929 2302.929

Year (Ref.: 2000) 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010

Number of observations −2*LL (starting values) −2*LL (final estimates) AIC BIC

69 1084133 1305563 1277925 1278037 1278704

0.927 0.917 0.903 0.971 1.103 1.170* 1.114 0.990 0.811** 0.507***

1.359*** 1.108 1.521*** 1.052

1.200

1084133 2162208 2103877 2103989 2104655

0.982 0.950 0.908 1.069 1.409*** 1.452*** 1.389*** 1.111 1.009 0.699**

1.059 1.381*** 2.387*** 1.407***

1.055

1.130 1.014 0.924

0.944

1.105 0.994

0.990***

Upward mobility

1084133 964066 850464 850574 851228

0.952 0.991 1.126** 1.223*** 1.580*** 1.756*** 1.613*** 1.392*** 1.339*** 1.050

1.159 1.292* 1.583*** 1.148*

1.144

1.032 0.954 0.878*

0.985

0.948 0.997

0.992***

Transition to fulltime employment

1084133 574364 536602 536712 537366

0.989 0.993 0.971 1.003 1.302*** 1.307*** 1.235** 1.247* 1.075 0.819

1.239 1.801*** 2.248*** 2.048***

1.466**

1.060 0.913 0.833*

0.986

1.020 0.952

0.990***

Transition to parttime employment

1084133 963832 906333 906443 907097

1.124* 1.420*** 2.003*** 2.421*** 3.114*** 3.148*** 2.943*** 2.710*** 2.235*** 1.465***

1.209* 1.101 1.285*** 0.927

1.212*

1.062 1.050 1.043

0.927

1.020 1.069

0.991***

Transition to marginal employment

1084133 7214303 6969041 6969153 6969820

1.149*** 1.318*** 1.378*** 1.195*** 0.882 0.813** 0.893 0.868 0.780* 1.082

2.132*** 1.030 1.262** 0.976

1.127

0.929 0.834** 0.735***

0.998

1.042 0.942

0.988***

Unemployment

Analyses are based on all employment periods of persons who took up employment between 1.1.2000 and 31.12.2010. The period of observation was from 1.1.2000 to 31.12.2010. *p < 0.05; **p < 0.01; ***p < 0.001. Hazard ratios for the Cox partial-likelihood estimates; Calculation of cluster-robust standard errors for 96 planning regions. Source: LIAB (LM 9310), own calculations.

1084133 1163982 1150572 1150684 1151350

0.981 0.915 0.977 1.026 1.040 1.006 0.947 0.840* 0.644*** 0.393***

0.978

0.953

0.960 1.052 0.945 0.828

1.059 0.902

0.989***

not bound by a collective agreement) 1.003 1.021 0.965 0.971

0.990***

0.993***

Downward mobility

Average gross daily wage of fulltime employees Collective agreement (Ref.: establishment Sectoral collective agreement Company collective agreement Orientation of sectoral collective agreement Firm size (Ref.: small firm) Small-medium-sized firm Medium-sized firm Larger firm Sector (Ref.: manufacturing industry) Agriculture, forestry, and mining Construction Trade Services for firms Other services

Lateral mobility

Exit from job

Independent variables

Table 14 (continued)

M. Dütsch, et al.

Advances in Life Course Research 40 (2019) 43–84

1.003 0.994 1.011 1.018 1.010 1.057 share of low-skilled workers) 1.006 0.997 1.009 0.957

1.004 0.993 1.013* 1.011 0.634 0.373 of low-skilled workers) 1.009 1.014 0.998 1.011 1.001 1.031* 1.012 0.967** 0.957**

0.958*** 0.967

0.950*** 0.942*

0.988 1.024** 1.158

0.989 0.989

1.029*** 1.015

1.006 1.007 0.996 1.018 1.937

0.981* 1.022** 0.543

70

0.044*** 1.015 0.005*** 0.001*** 0.056* 1.034*** 0.978

1.073*** 1.025 1.042** 1.048

0.020*** 5.828** 0.495 0.001*** 0.002*** 1.059*** 0.948** 1.007 1.079*** 1.070*** 1.040*** 0.940* 0.996 1.034 1.036

0.386 1.563 0.545 0.005*** 0.124 1.014 0.974 1.006 1.062*** 1.032 0.986 0.999 1.021 1.071** 0.968

Analyses are based on all employment periods of persons who took up employment between 1.1.2000 and 31.12.2010. The period of observation was from 1.1.2000 to 31.12.2010. Models additionally contain all explanatory variables that are reported in Table 3. *p < 0.05; **p < 0.01; ***p < 0.001. Hazard ratios for the Cox partial-likelihood estimates; Calculation of cluster-robust standard errors for 96 planning regions. Source: LIAB (LM 9310), own calculations.

1.010 0.803**

1.057**

1.054***

0.993 0.991

0.918*** 0.969*

1.043*** 1.028* 5.934***

0.969*** 0.993

1.023* 1.023* 3.643***

0.994 1.008

1.023* 1.019 1.429

Unemployment

0.978 0.992

1.046 1.045

1.011 1.009 0.396

0.980 1.013

1.017 1.012 1.831

1.028* 0.989

1.011 1.020* 0.982

Transition to marginal employment

0.981 1.039***

1.012 1.011

0.997 1.031 1.501

1.010 1.045*

0.996 1.025* 0.383

0.991 0.975

1.039*** 1.014

0.993 1.023 0.126**

0.981 1.000

1.002 1.047*** 6.559*

Transition to part-time employment

0.961*** 1.015 0.025***

Transition to full-time employment

0.987 1.011 0.654

0.988 0.996

0.971 0.993

1.003 1.024 2.828

Upward mobility

Human capital endowment (Ref.: share of low-skilled workers) Share of skilled workers 0.986* 0.989 0.967*** 0.978 0.942*** Share of highly skilled workers 1.000 0.989 0.991 0.991 1.010 Highest education level (Ref.: at most secondary school and no vocational training) *** *** At most secondary school and vocational training 0.068 0.613 0.142 0.278 0.012*** Advanced secondary school and no vocational training 3.399 0.771 0.083* 0.143 0.312 Advanced secondary school and vocational training 0.284* 0.556 0.022*** 0.144 0.014*** Polytechnic degree 0.196*** 0.330 0.004*** 0.178 0.009** University degree 0.714 0.107 0.004*** 0.042** 0.642 Highest education level (Ref.: at most secondary school and no vocational training) × human capital endowment (Ref.: share of low-skilled workers) *** 1.002 1.011 1.020 1.056*** At most secondary school and vocational training × share of 1.030 skilled workers Advanced secondary school and no vocational training × share 0.989 0.997 1.028 1.026 1.020 of skilled workers Advanced secondary school and vocational training × share of 1.011 0.997 1.038*** 1.023 1.057*** skilled workers Polytechnic degree × share of skilled workers 1.013* 1.007 1.053*** 1.026 1.047** University degree × share of skilled workers 1.006 1.029 1.065*** 1.049*** 1.029* 1.022 1.013 1.031 1.017 At most secondary school and vocational training × share of 1.025*** highly skilled workers Advanced secondary school and no vocational training × share 1.005 1.012 1.023 1.023 1.009 of highly skilled workers Advanced secondary school and vocational training × share of 1.023* 1.061** 1.050** 1.070** 1.018 highly skilled workers * Polytechnic degree × share of highly skilled workers 1.021 1.034 1.005 1.016 1.065* 0.991 0.998 1.028 0.979 University degree × share of highly skilled workers 0.911*

Human capital endowment (Ref.: share of low-skilled workers) Share of skilled workers Share of highly skilled workers Share of unemployment periods Share of unemployment periods × human capital endowment (Ref.: Share of unemployment periods × share of skilled workers Share of unemployment periods × share of highly skilled workers

Human capital endowment (Ref.: share of low-skilled workers) Share of skilled workers Share of highly skilled workers Nationality (1 = foreign) Nationality (1 = foreign) × human capital endowment (Ref.: share Foreign × share of skilled workers Foreign × share of highly skilled workers

0.992 1.013 2.658

Downward mobility

1.006 1.014 9.645**

Exit from job Lateral mobility

Human capital endowment (Ref.: share of low-skilled workers) Share of skilled workers 1.009 Share of highly skilled workers 1.011 Sex (1 = female) 1.947* Sex (1 = female) × human capital endowment (Ref.: share of low-skilled workers) Female × share of skilled workers 0.990* Female × share of highly skilled workers 1.000

Independent variables

Table 15 Estimates for cross-level interactions between regional human capital endowment and individual factors influencing employment trajectories with a redefined gap (more than 180 days without employment or registered unemployment).

M. Dütsch, et al.

Advances in Life Course Research 40 (2019) 43–84

1.020 1.145* 0.995

Region-specific factors Type of region (Ref.: rural area) Area in urbanisation process Urban area Unemployment rate 1.044 1.251** 0.993

Lateral mobility

71

1.540*** 1.110 0.800*** 0.473*** 1.977*** 1.259

0.861 1.251*** 1.469*** 1.460*** 0.418*** 0.330*** 0.562*** 0.597* 0.944 1.480* 0.974 1.147 0.705

0.675*** 0.586*** 0.413*** 0.424*** 0.768*** 5.106*** 1.696*** 0.809 1.494 3.188*** 1.604 1.630 0.844

4.378*** 1.336 0.822

2.896* 2.820* 0.887

1.119* 0.992 0.953 1.077* 0.925 1.082* 0.646*** 1.107 0.990***

0.979 1.038 1.025 0.874 1.013 0.675*** 1.011 0.991***

0.909 0.990***

*

0.662***

(continued on next page)

0.926 0.987***

0.977

0.948 1.064* 0.978 0.954

1.028

1.099**

1.120* 1.069

0.992 1.073 0.937 0.919

1.060*

1.028

0.910**

1.129 1.160 1.616*

1.027 0.411*** 0.769***

0.659*** 0.787***

0.709***

0.339***

0.934*

0.788*** 1.010* 0.999** 1.114**

1.016 0.915 1.015*

Unemployment

1.083 2.248 9.310***

1.032 –/– 0.602***

0.397*** 0.313***

0.640***

1.632***

0.757***

1.074* 0.974** 1.001*** 1.333***

1.112 1.089 0.959***

Transition to marginal employment

1.122**

2.555 3.077* 1.344

1.493 2.084 5.790***

–/– 1.225* 0.719***

0.829 1.104

0.753***

0.576***

0.903**

2.231*** 1.117*** 0.998*** 1.238***

1.067 1.335*** 0.988*

Transition to parttime employment

1.089*

0.938 0.868 0.542*

1.603* 1.988***

1.014

0.652***

1.050

1.017

0.719***

0.549*** 1.088*** 0.998*** 1.140**

1.148** 1.369*** 0.998

Transition to fulltime employment

0.836*** 1.009 0.999*** 0.948

1.074 1.247** 0.979***

Upward mobility

0.997 0.992 1.000** 1.014

1.132 1.027 0.999

Downward mobility

Firm-specific factors Employment prospects for the forthcoming year (Ref: will be approximately constant) The level of employment will 1.014 1.082 1.134** 1.179*** increase The level of employment will 1.072** 1.186** 1.189** 1.039 decrease Not sure at present 1.062* 1.074 0.985 1.021 Expected development of business volume in the current year compared to previous year (Ref.: it is expected to remain constant) * It is expected to increase 0.974 1.095 1.009 1.097* It is expected to decrease 1.057** 1.109* 1.106* 1.082 Do not know at present 0.986 1.298*** 0.936 0.876 Apprenticeships for vocational 0.926** 1.005 1.094 0.970 training are offered (1 = yes) Share of high-qualified 0.913* 0.925 0.660*** 0.919 employees * * Works council (1 = yes) 0.950 0.941 1.010 0.942 Average gross daily wage of full0.993*** 0.990*** 0.990*** 0.988*** time employees

Individual factors Sex (1 = female) 0.914*** 1.026 Age (in years) 0.975*** 1.030*** Age (in years squared) 1.000*** 1.000*** Nationality (1 = foreign) 1.196*** 0.859*** Highest education level (Ref.: at most secondary school and no vocational training) At most secondary school and 0.834*** 0.938 vocational training *** Advanced secondary school 1.328 0.675*** and no vocational training Advanced secondary school 0.868*** 0.836** and vocational training Polytechnic degree 0.773*** 0.879* University degree 0.911** 0.882* Employment state (Ref.: full-time) 0.544*** Part-time 0.923** Marginal employment 0.755*** 1.277* Daily wage (deflated) 0.727*** 1.418*** Previous employment trajectory (Ref.: previous vocational education) First employment 0.939 0.770 Share of full-time employment 1.363*** 1.197 *** Share of part-time 1.667 1.610* employment Share of unemployment 1.831*** 0.960 Share of nonemployment 1.566*** 1.091 Left-censored 1.035 0.768

Exit from job

Independent variables

Table 16 Estimates for individual, firm-specific and region-specific factors (focus on types of regions) influencing employment trajectories with redefined downward mobility (decrease in wages of more than 10%) and redefined upward mobility (increase in wages of at least 20%).

M. Dütsch, et al.

Advances in Life Course Research 40 (2019) 43–84

1.322

1.302***

1.731 1.191*** 1.402*** 1.162***

1.073*** 1.144*** 1.175*** 1.228*** 1.282*** 1.262*** 1.271*** 1.204*** 1.036 0.861***

1138538 12018907 11862703 2599.020 2599.020

Year (Ref.: 2000) 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010

Number of observations −2*LL (starting values) −2*LL (final estimates) AIC BIC

72

***

1208232 960285 939444 939554.1 940214.4

0.945 0.904 0.913 0.985 1.125 1.199* 1.141* 1.040 0.871* 0.572***

1.361 1.051 1.489*** 1.016

1.146

1208232 1461913 1411286 1411396 1412056

0.989 0.949 0.955 1.175* 1.561*** 1.642*** 1.551*** 1.259*** 1.170 0.850

0.994 1.338*** 2.650*** 1.458***

1.016

1.111 0.990 0.919

0.909

1.069 0.975

Upward mobility

1138538 8416876 7392044 1096.773 1098.257

0.948 0.990 1.126** 1.212*** 1.524*** 1.690*** 1.581*** 1.410*** 1.327*** 1.072

1.162 1.304** 1.662*** 1.185*

1.170

1.018 0.930 0.851**

0.969

0.931 0.981

Transition to fulltime employment

1138538 5112364 4763068 706.446 707.930

1.003 1.036 0.991 1.070 1.376*** 1.403*** 1.377*** 1.400*** 1.246*** 0.939

1.312 1.912*** 2.517*** 2.242***

1.460*

1.059 0.884 0.825*

0.975

1.004 0.931

Transition to parttime employment

1138538 405090 381730 763.569 764.217

1.087 1.399*** 2.071*** 2.071*** 2.393*** 2.927*** 2.833*** 2.588*** 2.201*** 1.495***

1.228 1.116 1.388*** 1.000

*

1.329*

1.051 1.055 1.033

0.922

1.005 1.070

Transition to marginal employment

1138538 3707672 3579912 715.993 716.059

1.145*** 1.298*** 1.344*** 1.155** 0.856* 0.785*** 0.889 0.869 0.783 1.034

2.215*** 1.057 1.303** 1.004

1.162

0.929 0.834** 0.731***

0.982

1.017 0.927

Unemployment

Analyses are based on all employment periods of persons who took up employment between 1.1.2000 and 31.12.2010. The period of observation was from 1.1.2000 to 31.12.2010. *p < 0.05; **p < 0.01; ***p < 0.001. Hazard ratios for the Cox partial-likelihood estimates; Calculation of cluster-robust standard errors for 96 planning regions. Source: LIAB (LM 9310), own calculations.

1208232 1962533 1938649 1938759 1939419

0.976 0.931 0.958 1.019 1.085 1.083 1.059 0.902 0.711*** 0.468***

1.632 1.638*** 2.296*** 1.673***

***

1.003 0.839 0.773*

1.075*** 1.034 0.977

***

0.980

0.979

0.958* 1.060 0.948 0.839

1.067 0.906

not bound by a collective agreement) 0.998 1.061 0.981 0.998

Downward mobility

Collective agreement (Ref.: establishment Sectoral collective agreement Company collective agreement Orientation of sectoral collective agreement Firm size (Ref.: small firm) Small-medium-sized firm Medium-sized firm Larger firm Sector (Ref.: manufacturing industry) Agriculture, forestry, and mining Construction Trade Services for firms Other services

Lateral mobility

Exit from job

Independent variables

Table 16 (continued)

M. Dütsch, et al.

Advances in Life Course Research 40 (2019) 43–84

1.032 1.236* 0.917 1.065 1.076

1.045 0.973

1.342 1.047

0.681* 0.649** 1.010 1.049 1.819***

1.033 1.255** 0.777

1.043 1.307***

1.092** 1.178*** 1.035 1.071* 1.710**

1.024 1.111 0.908

0.980 0.972 0.836***

Exit from job Lateral mobility

73 0.717 0.774* 0.748*** 0.644* 0.632** 1.136 0.627***

1.038 0.789* 0.171 0.782 1.474** 1.311** 0.797*

0.858 0.823 0.630**

0.733 0.703* 0.824** 0.729

0.746*

0.470

***

0.771 0.712* 0.893

0.739 0.761 0.768** 0.914

0.688*

0.667

*

0.736*

0.876 2.053*** 0.733** 0.531*** 0.397***

1.240 1.322*

1.153 0.887

1.078 1.103 2.858*

1.221 1.509*

1.109 1.059 0.971

0.912 0.742***

1.170* 1.274** 1.242***

Transition to marginal employment

0.773 0.464*** 0.911

1.029 1.096 0.922 1.001

0.884

0.862

0.816

0.957 0.510*** 0.778** 0.718*** 0.749**

1.019 0.969

1.169* 1.371***

0.958 0.829** 3.832***

1.047 1.126

1.016 0.908 1.026

1.044 1.097

0.998 0.882* 0.758***

Unemployment

Analyses are based on all employment periods of persons who took up employment between 1.1.2000 and 31.12.2010. The period of observation was from 1.1.2000 to 31.12.2010. Models additionally contain all explanatory variables that are reported in Table 2. *p < 0.05; **p < 0.01; ***p < 0.001. Hazard ratios for the Cox partial-likelihood estimates; Calculation of cluster-robust standard errors for 96 planning regions. Source: LIAB (LM 9310), own calculations.

0.668**

0.667 0.712*

0.571

*

0.821*

0.797**

1.128 *

1.058 0.901 0.882 0.997 1.544**

1.095 0.941 1.141 1.746*** 2.691***

1.313*** 1.584***

1.179 1.206

1.225 0.916 1.243* 1.369** 1.761***

1.088 1.172

1.025 1.282*** 2.267

1.074 1.065

1.064 1.332*** 1.166

0.918 0.916

1.133 1.419*** 2.382***

Transition to part-time employment

1.417*** 1.777***

1.277 1.251

1.396** 1.525**

1.114 1.271*** 0.804

0.890 0.884

1.151* 1.375*** 1.275

0.998 0.877

1.151* 1.470*** 1.088***

Transition to full-time employment

1.042 1.120

1.017 1.189* 0.831

1.028 0.850

1.060 1.242** 1.070

1.011 1.300***

0.821* 1.214* 0.746***

Upward mobility

1.136 1.230 1.056

1.003 0.957

1.130 1.231** 1.041

1.102 1.101

1.082 1.175* 0.927

Downward mobility

Type of region (Ref.: rural area) Area in urbanisation process 1.124 1.231 1.258 ** Urban area 1.132* 1.318** 1.319** Highest education level (Ref.: at most secondary school and no vocational training) At most secondary school and vocational training 0.903* 1.024 0.810* Advanced secondary school and no vocational training 1.509*** 1.319* 0.955 Advanced secondary school and vocational training 0.882* 0.981 0.641*** Polytechnic degree 0.807*** 0.954 0.529*** University degree 0.972 1.137 0.575*** Highest education level (Ref.: at most secondary school and no vocational training) × type of region (Ref.: rural area) At most secondary school and vocational training × area 0.897 0.718* 0.920 undergoing urbanisation process Advanced secondary school and no vocational training × area 0.831 1.024 0.369*** undergoing urbanisation process Advanced secondary school and vocational training × area 0.915 0.858 0.531** undergoing urbanisation process Polytechnic degree × area undergoing urbanisation process 0.857* 0.793* 0.653** University degree × area undergoing urbanisation process 0.912 0.746* 0.789 At most secondary school and vocational training × urban area 0.895* 0.989 0.970 Advanced secondary school and no vocational training × urban 0.878 0.867 0.893 area Advanced secondary school and vocational training × urban area 1.012 0.623** 0.624** Polytechnic degree × urban area 1.001 0.486** 0.767* University degree × urban area 0.919 0.971 0.624***

Type of region (Ref.: rural area) Area in urbanisation process Urban area Share of unemployment periods Share of unemployment periods × type of region (Ref.: rural area) Share of unemployment periods × area in urbanisation process Share of unemployment periods × urban area

Type of region (Ref.: rural area) Area in urbanisation process Urban area Nationality (1 = foreign) Nationality (1 = foreign) × type of region (Ref.: rural area) Foreign × area in urbanisation process Foreign × urban area

Type of region (Ref.: rural area) Area in urbanisation process Urban area Sex (1 = female) Sex (1 = female) × type of region (Ref.: rural area) Female × area in urbanisation process Female × urban area

Independent variables

Table 17 Estimates for cross-level interactions between types of regions and individual factors influencing employment trajectories with redefined downward mobility (decrease in wages of more than 10%) and redefined upward mobility (increase in wages of at least 20%).

M. Dütsch, et al.

Advances in Life Course Research 40 (2019) 43–84

Exit from job

74 0.477** 1.985** 1.279

1.264*** 1.500*** 1.500*** 0.418*** 0.330*** 0.563*** 0.600* 0.949 1.488* 0.982 1.152 0.709

0.596*** 0.424*** 0.441*** 0.769*** 5.006*** 1.677*** 0.825 1.526 3.245*** 1.642 1.652 0.856

Firm-specific factors Employment prospects for the forthcoming year (Ref: will be approximately constant) The level of employment will 1.010 1.070 1.117* 1.160** increase The level of employment will 1.069** 1.183** 1.180** 1.034 decrease ** Not sure at present 1.061 1.076 0.982 1.017 Expected development of business volume in the current year compared to previous year (Ref.: it is expected to remain constant) * It is expected to increase 0.974 1.094 1.012 1.095* 1.107* 1.105* 1.079 It is expected to decrease 1.058** Do not know at present 0.986 1.285** 0.941 0.871 1.004 1.097 0.970 Apprenticeships for vocational 0.928** training are offered (1 = yes) Share of high-qualified 0.950* 0.948 0.683*** 0.929 employees Works council (1 = yes) 0.912* 0.942 1.006 0.940

2.643*** 1.109 0.801***

0.889

0.713***

4.295*** 1.322 0.813

2.811* 2.785* 0.866

1.120* 0.979 0.953 1.071 0.923 1.083 0.659*** 1.089

0.976 1.032 1.024 0.882 1.014 0.675*** 1.000

0.929

0.906*

(continued on next page)

0.963

0.646***

0.947 1.064* 0.990 0.962

1.034

1.103**

1.105* 1.062

0.998 1.074 0.944 0.927

1.058*

1.026

0.906**

1.113 1.023 2.039***

1.026 0.416*** 0.771***

0.648*** 0.777***

0.701***

0.337***

0.927*

0.788*** 1.011* 1.000*** 1.103**

1.002

1.021* 1.024*

Unemployment

1.066 2.211 9.209***

1.028 –/– 0.602***

0.393*** 0.313***

0.635***

1.638***

0.747***

1.071* 0.975*** 1.000** 1.342***

0.946***

1.020 1.019*

Transition to marginal employment

1.096*

2.606 3.115* 1.369

1.518 2.112 5.886***

–/– 1.215* 0.718***

0.854 1.141

0.768***

0.617***

0.915**

2.214*** 1.116*** 0.998*** 1.247***

0.980**

1.002 1.034**

Transition to parttime employment

1.084*

0.945 0.879 0.547*

1.605* 2.035***

1.024

0.673***

1.051

1.029

0.730***

0.546*** 1.088*** 0.998*** 1.143*

0.998

0.972** 0.832*** 1.009 0.999*** 0.949

0.999

0.995

0.980* 1.022**

Transition to fulltime employment

1.003 1.027

Upward mobility

0.992 0.991 1.000** 1.010

0.990 1.011

Downward mobility

0.987 1.013

Lateral mobility

Individual factors Sex (1 = female) 0.913*** 1.022 Age (in years) 0.975*** 1.029*** Age (in years squared) 1.000*** 0.999*** Nationality (1 = foreign) 1.196*** 0.859*** Highest education level (Ref.: at most secondary school and no vocational training) 0.951 At most secondary school and 0.835*** vocational training Advanced secondary school 1.340*** 0.696*** and no vocational training Advanced secondary school 0.869*** 0.847** and vocational training Polytechnic degree 0.774*** 0.900* University degree 0.914* 0.911 Employment state (Ref.: full-time) Part-time 0.924** 0.544*** 1.246* Marginal employment 0.757*** Daily wage (deflated) 0.727*** 1.403*** Previous employment trajectory (Ref.: previous vocational education) First employment 0.938 0.782 Share of full-time employment 1.361*** 1.215 *** 1.631* Share of part-time 1.667 employment Share of unemployment 1.829*** 0.975 Share of nonemployment 1.567*** 1.099 Left-censored 1.034 0.775

Region-specific factors Human capital endowment (Ref.: share of low-skilled workers) Share of skilled workers 1.003 Share of highly skilled 1.009 workers Unemployment rate 0.995

Independent variables

Table 18 Estimates for individual, firm-specific and region-specific factors (focus on human capital endowment) influencing employment trajectories with redefined downward mobility (decrease in wages of more than 10%) and redefined upward mobility (increase in wages of at least 20%).

M. Dütsch, et al.

Advances in Life Course Research 40 (2019) 43–84

0.998 0.852* 0.766** 1.352 1.588*** 1.612*** 2.219*** 1.621***

1.072** 1.035 0.971

1.308***

1.721*** 1.184*** 1.388*** 1.155***

1.067*** 1.135*** 1.163*** 1.208*** 1.259*** 1.228*** 1.223*** 1.149*** 0.983 0.812***

1138538 12018907 11862703 2599.020 2599.020

Year (Ref.: 2000) 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010

Number of observations −2*LL (starting values) −2*LL (final estimates) AIC BIC

75 1138538 960285 939242 939354 940027

0.931 0.883* 0.877 0.922 1.035 1.083 1.017 0.911 0.757*** 0.492***

1.323** 1.024 1.427*** 0.990

1.171

1138538 1461913 1410939 1411051 1411723

0.969 0.917 0.908 1.086 1.415*** 1.446*** 1.327* 1.042 0.956 0.682*

0.966 1.310*** 2.555*** 1.420***

1.035

1.102 0.999 0.900

0.912

1.088 0.963

0.988***

Upward mobility

1138538 420843 369622 739355 740015

0.942 0.973 1.097* 1.179*** 1.464*** 1.613*** 1.498*** 1.321*** 1.232* 0.996

1.153 1.306*** 1.635*** 1.191*

1.188

1.014 0.932 0.854**

0.974

0.949 0.971

0.991***

Transition to fulltime employment

1138538 511236 476199 476309 476969

0.978 0.993 0.931 0.965 1.219** 1.200** 1.315*** 1.111 0.970 0.720**

1.258 1.872*** 2.359*** 2.180***

1.484*

1.048 0.893 0.798**

0.979

1.021 0.919

0.990***

Transition to parttime employment

1138538 810182 763464 763574 764235

1.078 1.387*** 2.068*** 2.375*** 2.850*** 2.865*** 2.689*** 2.401*** 2.018*** 1.348***

1.221* 1.117 1.381*** 1.001

1.319*

1.046 1.044 1.019

0.932

1.020 1.073

0.990***

Transition to marginal employment

1138538 7415344 7159576 7159688 7160360

1.144*** 1.297*** 1.355*** 1.165** 0.868* 0.782*** 0.856 0.828 0.737** 0.965

2.197*** 1.042 1.295*** 0.994

1.148

0.926 0.827** 0.717***

0.995

1.035 0.938

0.987***

Unemployment

Analyses are based on all employment periods of persons who took up employment between 1.1.2000 and 31.12.2010. The period of observation was from 1.1.2000 to 31.12.2010. *p < 0.05; **p < 0.01; ***p < 0.001. Hazard ratios for the Cox partial-likelihood estimates; Calculation of cluster-robust standard errors for 96 planning regions. Source: LIAB (LM 9310), own calculations.

1138538 1962533 1938189 1938301 1938973

0.963 0.907 0.914 0.947 0.986 0.968 0.935 0.786** 0.614*** 0.398***

0.990

0.980

0.961 1.061 0.965 0.834

1.087 0.898

0.990***

not bound by a collective agreement) 1.006 1.073 0.981 0.981

0.989***

0.993***

Downward mobility

Average gross daily wage of fulltime employees Collective agreement (Ref.: establishment Sectoral collective agreement Company collective agreement Orientation of sectoral collective agreement Firm size (Ref.: small firm) Small-medium-sized firm Medium-sized firm Larger firm Sector (Ref.: manufacturing industry) Agriculture, forestry, and mining Construction Trade Services for firms Other services

Lateral mobility

Exit from job

Independent variables

Table 18 (continued)

M. Dütsch, et al.

Advances in Life Course Research 40 (2019) 43–84

1.003 0.991 1.008 1.021 0.969 3.490 share of low-skilled workers) 1.001 0.977 1.011 0.946

1.002 0.987 1.011* 1.014 0.519 0.359 of low-skilled workers) 1.013 1.015 0.981 0.987 1.010 1.034* 1.008 0.959** 0.956**

0.947*** 0.963

0.939*** 0.958*

0.989 1.027*** 1.264

0.990 0.992

1.026** 1.007

1.009 1.005 1.000 1.018 2.811**

0.981* 1.022** 0.481

1.042*** 1.017

1.002 1.027 0.170*

0.984 1.006

0.960*** 1.014 0.018***

Transition to full-time employment

0.989 1.011 0.527

0.991 0.996

0.974** 0.997

1.010 1.025 1.938

Upward mobility

76

0.045*** 0.717 0.006*** 0.002*** 0.065* 1.034*** 0.981

1.069*** 1.029 1.040** 1.061

0.019*** 6.598*** 0.335 0.001*** 0.005*** 1.051*** 0.959** 1.012 1.083*** 1.063*** 1.031* 0.942* 0.994 1.018 1.027

0.331 1.770 0.447 0.005*** 0.197 1.017 0.971 1.009 1.061*** 1.033 0.984 1.026 1.028 1.082** 0.974

1.010 0.810**

1.055**

1.052***

0.994 0.994

0.920*** 0.968*

1.042*** 1.029* 5.105***

0.969*** 0.986

1.023* 1.024* 3.123***

0.993 1.008

1.023* 1.020 1.441

Unemployment

0.984 1.001

1.021 1.029

1.018 1.015 0.377

0.972 1.001

1.020 1.016 1.890

1.021* 0.994

1.008 1.022* 0.363

Transition to marginal employment

0.988 1.040**

0.998 0.9998

1.004 1.035 1.935

1.012 1.037*

1.001 1.030** 0.342

0.992 0.971

1.006 1.053*** 6.230***

Transition to part-time employment

Analyses are based on all employment periods of persons who took up employment between 1.1.2000 and 31.12.2010. The period of observation was from 1.1.2000 to 31.12.2010. Models additionally contain all explanatory variables that are reported in Table 3. *p < 0.05; **p < 0.01; ***p < 0.001. Hazard ratios for the Cox partial-likelihood estimates; Calculation of cluster-robust standard errors for 96 planning regions. Source: LIAB (LM 9310), own calculations.

Human capital endowment (Ref.: share of low-skilled workers) Share of skilled workers 0.987* 0.984 0.972** 0.974 0.941*** Share of highly skilled workers 0.997 0.990 0.996 0.994 1.006 Highest education level (Ref.: at most secondary school and no vocational training) *** * At most secondary school and vocational training 0.098 0.906 0.239 0.240 0.014*** Advanced secondary school and no vocational training 4.758 0.426 0.081* 0.097 0.222 Advanced secondary school and vocational training 0.259* 0.789 0.017*** 0.125 0.007** Polytechnic degree 0.161*** 0.642 0.007*** 0.101 0.006** University degree 0.470 0.107 0.002*** 0.040* 0.611 Highest education level (Ref.: at most secondary school and no vocational training) × human capital endowment (Ref.: share of low-skilled workers) *** * 0.999 1.012 1.028 1.056*** At most secondary school and vocational training × share of 1.026 skilled workers Advanced secondary school and no vocational training × share 0.989 1.004 1.027 1.032 1.018 of skilled workers Advanced secondary school and vocational training × share of 1.013 0.995 1.042*** 1.026 1.065*** skilled workers Polytechnic degree × share of skilled workers 1.016** 1.002 1.046*** 1.034 1.051** University degree × share of skilled workers 1.011 1.030 1.073*** 1.051** 1.029* 1.022 1.006 1.032 1.021 At most secondary school and vocational training × share of 1.019*** highly skilled workers Advanced secondary school and no vocational training × share 1.001 1.012 1.080 1.094 1.028 of highly skilled workers Advanced secondary school and vocational training × share of 1.023** 1.056** 1.055** 1.071* 1.026 highly skilled workers * Polytechnic degree × share of highly skilled workers 1.020 1.023 1.007 1.019 1.080* 1.006 0.995 1.028 0.983 University degree × share of highly skilled workers 0.912**

Human capital endowment (Ref.: share of low-skilled workers) Share of skilled workers Share of highly skilled workers Share of unemployment periods Share of unemployment periods × human capital endowment (Ref.: Share of unemployment periods × share of skilled workers Share of unemployment periods × share of highly skilled workers

Human capital endowment (Ref.: share of low-skilled workers) Share of skilled workers Share of highly skilled workers Nationality (1 = foreign) Nationality (1 = foreign) × human capital endowment (Ref.: share Foreign × share of skilled workers Foreign × share of highly skilled workers

0.994 1.013 2.277

Downward mobility

0.999 1.014 6.549**

Exit from job Lateral mobility

Human capital endowment (Ref.: share of low-skilled workers) Share of skilled workers 1.010 Share of highly skilled workers 1.010 Sex (1 = female) 2.135* Sex (1 = female) × human capital endowment (Ref.: share of low-skilled workers) Female × share of skilled workers 0.988** Female × share of highly skilled workers 1.004

Independent variables

Table 19 Estimates for cross-level interactions between regional human capital endowment and individual factors influencing employment trajectories with redefined downward mobility (decrease in wages of more than 10%) and redefined upward mobility (increase in wages of at least 20%).

M. Dütsch, et al.

Advances in Life Course Research 40 (2019) 43–84

1.019 1.147* 0.999

Region-specific factors Type of region (Ref.: rural area) Area in urbanisation process Urban area Unemployment rate 1.043 1.220* 0.996

Lateral mobility

77

1.610*** 1.099 0.794*** 0.473*** 1.980*** 1.260

0.853 1.211*** 1.574*** 1.811*** 0.380*** 0.363*** 0.631*** 0.566*** 0.985 1.292 0.853 0.976 0.660**

0.680*** 0.598*** 0.494*** 0.539*** 0.745*** 6.792*** 2.159*** 0.745 1.017 1.859** 1.021 1.012 0.814

4.394*** 1.339 0.826

2.873* 2.797* 0.886

1.117* 0.993 0.954 1.077* 0.926 1.083* 0.646*** 1.109 0.990***

0.979 1.038 1.024 0.874 1.013 0.676*** 1.011 0.991***

0.926 0.988***

0.910* 0.991***

(continued on next page)

0.978

0.665***

0.949 1.064* 0.978 0.954

1.028

1.100**

1.121* 1.069

0.994 1.074 0.939 0.916

1.061*

1.030

0.910**

1.135 1.171 1.775*

1.020 0.395*** 0.754***

0.669*** 0.805***

0.715***

0.340***

0.936*

0.787*** 1.011* 1.000*** 1.112**

1.015 0.915 1.015**

Unemployment

1.084 2.260 9.278***

1.011 –/– 0.572***

0.411*** 0.328***

0.652***

1.621***

0.767***

1.065* 0.976** 1.000** 1.332***

1.110 1.092 0.960***

Transition to marginal employment

1.123**

2.547 3.079* 1.346

1.493 2.099 5.799***

–/– 1.279* 0.705***

0.836 1.127

0.759***

0.577***

0.904**

2.225*** 1.118*** 0.998*** 1.237***

1.065 1.335*** 0.988*

Transition to parttime employment

1.090*

0.937 0.868 0.542*

1.604* 1.991***

1.015

0.652***

1.050

0.968

0.709***

0.549*** 1.088*** 0.998*** 1.140**

1.148** 1.371*** 0.998

Transition to fulltime employment

0.846*** 1.029*** 0.999*** 0.899

1.062 1.295*** 0.984**

Upward mobility

0.978 1.005 1.000*** 0.963

1.128 1.019 0.999

Downward mobility

Firm-specific factors Employment prospects for the forthcoming year (Ref: will be approximately constant) The level of employment will 1.015 1.040 1.116** 1.167*** increase The level of employment will 1.072** 1.232** 1.227*** 1.100 decrease Not sure at present 1.062* 1.129 1.034 1.085 Expected development of business volume in the current year compared to previous year (Ref.: it is expected to remain constant) It is expected to increase 0.974 1.100* 0.997 1.084* It is expected to decrease 1.058** 1.114* 1.098* 1.086* Do not know at present 0.986 1.331*** 1.021 0.956 Apprenticeships for vocational 0.926** 0.937 1.025 0.928 training are offered (1 = yes) Share of high-qualified 0.912* 0.963 0.696** 0.905 employees Works council (1 = yes) 0.950 0.893* 0.969 0.934 0.992*** 0.991*** 0.991*** Average gross daily wage of full0.993*** time employees

Individual factors Sex (1 = female) 0.913*** 0.986 Age (in years) 0.976*** 1.047*** *** Age (in years squared) 1.000 0.999*** Nationality (1 = foreign) 1.193*** 0.860** Highest education level (Ref.: at most secondary school and no vocational training) At most secondary school and 0.835*** 0.961 vocational training *** Advanced secondary school 1.331 0.711** and no vocational training Advanced secondary school 0.872*** 0.806*** and vocational training Polytechnic degree 0.783*** 0.895* University degree 0.929* 0.894* Employment state (Ref.: full-time) Part-time 0.917*** 0.602*** Marginal employment 0.739*** 1.734*** Daily wage (deflated) 0.719*** 1.608*** Previous employment trajectory (Ref.: previous vocational education) First employment 0.944 0.918 Share of full-time employment 1.377*** 0.742 Share of part-time 1.675*** 1.444 employment *** Share of unemployment 1.836 1.105 Share of nonemployment 1.571*** 1.159 Left-censored 1.040 0.902

Exit from job

Independent variables

Table 20 Estimates for the individual, firm-specific and region-specific factors (focus on types of regions) influencing employment trajectories using imputed right censored wages.

M. Dütsch, et al.

Advances in Life Course Research 40 (2019) 43–84

1.233

1.298***

1.727 1.191*** 1.398*** 1.158***

1.073*** 1.144*** 1.174*** 1.226*** 1.281*** 1.261*** 1.269*** 1.204*** 1.035 0.859***

1208113 24034704 23721699 23700000 23700000

Year (Ref.: 2000) 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010

Number of observations −2*LL (starting values) −2*LL (final estimates) AIC BIC

78

***

1208113 1370932 1347549 1347659 1348319

0.965 0.875* 0.912 0.936 1.094 1.108 1.049 0.903 0.745*** 0.492

1.398 1.206* 1.621*** 1.061

1.235

1208113 2419757 2368457 2368567 2369227

0.987 0.962 0.872 1.017 1.344*** 1.378*** 1.331*** 1.104 1.001 0.701***

0.994 1.365*** 2.123*** 1.258***

0.923

1.148* 1.037 0.920

0.934

1.134 1.023

Upward mobility

1208113 841578 739084 739192 739840

0.949 0.991 1.126** 1.212*** 1.525*** 1.690*** 1.582*** 1.410*** 1.329*** 1.074

1.162 1.303** 1.660*** 1.183*

1.169

1.018 0.930 0.851**

0.968

0.931 0.982

Transition to fulltime employment

1208113 511107 476132 476242 476902

1.004 1.037 0.991 1.070 1.376*** 1.402*** 1.377*** 1.401*** 1.247*** 0.939

1.310 1.914*** 2.510*** 2.232***

1.455*

1.057 0.883 0.825*

0.975

1.004 0.930

Transition to parttime employment

1208113 810100 763050 763158 763806

1.084 1.395*** 2.056*** 2.376*** 2.829*** 2.910*** 2.818*** 2.593*** 2.193*** 1.490***

1.229 1.116 1.383*** 0.994

*

1.318*

1.050 1.054 1.030

0.921

1.007 1.074

Transition to marginal employment

1208113 7414805 7158752 7158862 7159522

1.146*** 1.299*** 1.344*** 1.154*** 0.856** 0.785*** 0.883 0.870 0.784** 1.035

2.211*** 1.056 1.298*** 1.001

1.158

0.930 0.835** 0.732***

0.982

1.018 0.928

Unemployment

Analyses are based on all employment periods of persons who took up employment between 1.1.2000 and 31.12.2010. The period of observation was from 1.1.2000 to 31.12.2010. *p < 0.05; **p < 0.01; ***p < 0.001. Hazard ratios for the Cox partial-likelihood estimates; Calculation of cluster-robust standard errors for 96 planning regions. Source: LIAB (LM 9310), own calculations.

1208113 1385110 1369620 1369730 1370390

0.962 0.842* 0.998 1.055 1.107 0.996 0.979 0.884 0.666*** 0.407***

1.553 1.743*** 1.963*** 1.459***

***

0.948 0.811** 0.718***

1.075** 1.034 0.977

***

0.914

0.900

0.958* 1.072 0.956 0.826

0.963 0.841

not bound by a collective agreement) 0.999 0.961 0.982 0.959

Downward mobility

Collective agreement (Ref.: establishment Sectoral collective agreement Company collective agreement Orientation of sectoral collective agreement Firm size (Ref.: small firm) Small-medium-sized firm Medium-sized firm Larger firm Sector (Ref.: manufacturing industry) Agriculture, forestry, and mining Construction Trade Services for firms Other services

Lateral mobility

Exit from job

Independent variables

Table 20 (continued)

M. Dütsch, et al.

Advances in Life Course Research 40 (2019) 43–84

1.042 1.231* 1.130 1.015 0.930

1.044 0.969

1.253 1.042

0.681* 0.648** 1.010 1.050 1.828***

1.035 1.222* 0.795

1.037 1.202*

1.091* 1.173*** 1.034 1.072* 1.707***

1.027 1.129 0.897

0.980 0.975 0.837***

Exit from job Lateral mobility

79 0.538** 1.025 0.707**

1.356** 1.333* 0.726*

0.860 0.823 0.632**

0.732 0.700* 0.822** 0.726

0.748

*

0.911 0.568** 1.032

0.884 1.091 0.922 0.771

0.688* 0.753 0.771* 0.764 0.885 0.672* 0.739

0.815

0.735

0.864

1.008

0.961 0.511*** 0.784** 0.727*** 0.767*

1.020 0.970

1.169* 1.368***

0.958 0.830** 3.849***

1.047 1.125

1.015 0.909 1.025

1.042 1.092

0.998 0.884* 0.759***

Unemployment

*

0.705*

0.811*

0.888 2.058*** 0.751** 0.549*** 0.418***

1.241 1.323*

1.157 0.890

1.076 1.104 2.831*

1.218 1.508*

1.107 1.062 0.972

0.905 0.733***

1.172* 1.286** 1.242***

Transition to marginal employment

Analyses are based on all employment periods of persons who took up employment between 1.1.2000 and 31.12.2010. The period of observation was from 1.1.2000 to 31.12.2010. Models additionally contain all explanatory variables that are reported in Table 2. *p < 0.05; **p < 0.01; ***p < 0.001. Hazard ratios for the Cox partial-likelihood estimates; Calculation of cluster-robust standard errors for 96 planning regions. Source: LIAB (LM 9310), own calculations.

0.802 0.711* 8.005* 0.712*

0.615 1.017 0.672* 1.200 0.850

0.701

**

0.516*

0.532* *

0.818*

0.790**

1.025

0.468***

1.061 0.906 0.889 1.007 1.576***

0.936 1.059 0.914 1.520*** 2.685***

1.315** 1.586***

1.177 1.202

0.887 1.184 1.153 1.356*** 2.255***

1.065 1.153

1.024 1.283*** 2.265

1.078 1.065

1.062 1.332*** 1.164

0.917 0.911

1.133 1.424*** 2.381***

Transition to part-time employment

1.270** 1.255*

1.255 1.142

1.343** 1.375*

1.123 1.375* 0.823

1.054 0.966

1.164* 1.223** 0.972

1.053 1.095

1.153* 1.173* 1.092***

Transition to full-time employment

1.124 1.134

1.009 1.257** 0.748

0.929 0.758

1.062 1.314*** 1.094

1.045 1.210**

0.843* 1.198* 0.767***

Upward mobility

1.065 1.153 0.823

1.054 0.966

1.124 1.223** 0.972

1.053 1.095

1.103 0.194* 0.925

Downward mobility

Type of region (Ref.: Rural area) Area in urbanisation process 1.125 1.235 1.270** Urban area 1.132* 1.232* 1.255* Highest education level (Ref.: at most secondary school and no vocational training) At most secondary school and vocational training 0.904* 0.902 0.736*** Advanced secondary school and no vocational training 1.512*** 1.369 1.059 Advanced secondary school and vocational training 0.886* 0.867 0.614*** *** * Polytechnic degree 0.816 0.828 0.520*** University degree 0.992 0.997 0.685*** Highest education level (Ref.: at most secondary school and no vocational training) × types of region (Ref.: rural area) 0.900 At most secondary school and vocational training × area 0.896 0.630** undergoing urbanisation process Advanced secondary school and no vocational training × area 0.831 1.082 0.516* undergoing urbanisation process Advanced secondary school and vocational training × area 0.915 0.741 0.615* undergoing urbanisation process Polytechnic degree × area undergoing urbanisation process 0.856* 0.788* 0.702** University degree × area undergoing urbanisation process 0.907 0.923 0.810 At most secondary school and vocational training × urban area 0.895* 1.018 1.005 Advanced secondary school and no vocational training × urban 0.878 0.871 0.912 area * Advanced secondary school and vocational training × urban area 1.013 0.662 0.581** Polytechnic degree × urban area 1.009 0.552** 0.625* University degree × urban area 0.923 0.937 0.907

Type of region (Ref.: rural area) Area in urbanisation process Urban area Share of unemployment periods Share of unemployment periods × Types of region (Ref.: rural area) Share of unemployment periods × area in urbanisation process Share of unemployment periods × urban area

Type of region (Ref.: Rural area) Area in urbanisation process Urban area Nationality (1 = foreign) Nationality (1 = foreign) × types of region (Ref.: rural area) Foreign × area in urbanisation process Foreign × urban area

Type of region (Ref.: Rural area) Area in urbanisation process Urban area Sex (1 = female) Sex (1 = female) × types of region (Ref.: rural area) Female × area in urbanisation process Female × urban area

Independent variables

Table 21 Estimates for cross-level interactions between types of regions and individual factors influencing employment trajectories using imputed right censored wages.

M. Dütsch, et al.

Advances in Life Course Research 40 (2019) 43–84

Exit from job

80 0.477** 1.988** 1.280

1.222*** 1.604*** 1.855*** 0.380*** 0.362*** 0.631*** 0.571*** 0.991 1.504* 0.858 0.981 0.663**

0.606*** 0.508*** 0.562*** 0.746*** 6.616*** 2.130*** 0.856 1.040 1.891** 1.039 1.024 0.520**

Firm-specific factors Employment prospects for the forthcoming year (Ref: will be approximately constant) The level of employment will 1.011 1.026 1.094* 1.148** increase The level of employment will 1.070** 1.226** 1.216*** 1.095 decrease Not sure at present 1.061* 1.129 1.029 1.081 Expected development of business volume in the current year compared to previous year (Ref.: it is expected to remain constant) It is expected to increase 0.974 1.101* 1.000 1.081* 1.114* 1.098* 1.084* It is expected to decrease 1.058** Do not know at present 0.985 1.312*** 1.021 0.945 0.936 1.028 0.928 Apprenticeships for vocational 0.927** training are offered (1 = yes) Share of high-qualified 0.911* 0.977 0.717** 0.911 employees * ** Works council (1 = yes) 0.949 0.895 0.967 0.933 Average gross daily wage of full0.993*** 0.991*** 0.990*** 0.991*** time employees

1.923*** 1.091 0.795***

0.883

0.721**

4.311*** 1.325 0.817

2.792* 2.764* 0.866

1.118* 0.979 0.954 1.072* 0.923 1.084* 0.659*** 1.090 0.990***

0.976 1.032 1.024 0.882 1.014 0.675*** 1.000 0.991***

0.907 0.991***

*

0.649***

(continued on next page)

0.928 0.988***

0.963

0.948 1.064* 0.989 0.962

1.034

1.104**

1.106* 1.063

1.000 1.075 0.946 0.924

1.058*

1.029

0.907**

1.118 1.044 2.048***

1.019 0.399*** 0.756***

0.659*** 0.795***

0.706***

0.337***

0.929*

0.787*** 1.012** 1.000*** 1.101*

1.002

1.021* 1.024*

Unemployment

1.067 2.225 9.182***

1.008 –/– 0.572***

0.407*** 0.327***

0.647***

1.628***

0.757***

1.062* 0.977** 1.000** 1.341***

0.946***

1.020 1.019*

Transition to marginal employment

1.096*

2.599 3.117* 1.371

1.518 2.128 5.895***

–/– 1.196* 0.703***

0.862 1.146

0.775***

0.618***

0.915**

2.208*** 1.117*** 0.998*** 1.247***

0.979**

1.002 1.034**

Transition to parttime employment

1.085*

0.944 0.879 0.547*

1.606* 2.038***

1.025

0.673***

1.051

0.979

0.720***

0.547*** 1.088*** 0.998*** 1.143**

0.998

0.975* 0.842*** 1.029*** 0.999*** 0.923

0.993

0.992

0.980* 1.029**

Transition to fulltime employment

1.004 1.023

Upward mobility

0.973 1.004 1.000*** 0.962

1.001 1.017

Downward mobility

1.001 1.017

Lateral mobility

Individual factors Sex (1 = female) 0.912*** 0.984 Age (in years) 0.976*** 1.046*** Age (in years squared) 1.000*** 0.999*** Nationality (1 = foreign) 1.193*** 0.864** Highest education level (Ref.: at most secondary school and no vocational training) 0.971 At most secondary school and 0.836*** vocational training *** Advanced secondary school 1.343 0.731* and no vocational training Advanced secondary school 0.873*** 0.813*** and vocational training Polytechnic degree 0.784*** 0.915 University degree 0.923* 0.973 Employment state (Ref.: full-time) *** Part-time 0.918 0.602*** 1.695*** Marginal employment 0.741*** Daily wage (deflated) 0.720*** 1.596*** Previous employment trajectory (Ref.: previous vocational education) First employment 0.943 0.932 Share of full-time employment 1.376*** 1.467 1.763* Share of part-time 1.675*** employment Share of unemployment 1.835*** 1.116 Share of nonemployment 1.572*** 1.167 Left-censored 1.039 0.908

Region-specific factors Human capital endowment (Ref.: share of low-skilled workers) Share of skilled workers 1.004 Share of highly skilled 1.009 workers Unemployment rate 0.995

Independent variables

Table 22 Estimates for the individual, firm-specific and region-specific factors (focus on human capital endowment) influencing employment trajectories using imputed right censored wages.

M. Dütsch, et al.

Advances in Life Course Research 40 (2019) 43–84

1.256

1.304***

1.716 1.184*** 1.383*** 1.151***

1.067*** 1.135*** 1.162*** 1.205*** 1.256*** 1.225*** 1.220*** 1.147*** 0.980 0.809***

1208113 24034704 23721345 2599.020 2599.020

Year (Ref.: 2000) 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010

Number of observations −2*LL (starting values) −2*LL (final estimates) AIC BIC

81

***

1208113 1370932 1347176 1347288 1347960

0.945 0.849** 0.870 0.862* 0.992 0.975 0.898 0.752*** 0.612*** 0.397***

1.353 1.170 1.544*** 1.027

1.261

1208113 2419757 2367747 2367859 2368531

0.965 0.926 0.824* 0.932 1.206* 1.194 1.112 0.890 0.793 0.544***

0.962 1.337*** 2.038*** 1.219**

0.945

1.037* 1.048 0.900

0.937

1.154* 1.007

Upward mobility

1208113 841578 739125 739235 739895

0.942 0.974 1.097* 1.178*** 1.464*** 1.613*** 1.498*** 1.321*** 1.232* 0.996

1.153 1.306** 1.633*** 1.189*

1.187

1.014 0.932 0.854**

0.973

0.949 0.972

Transition to fulltime employment

1208113 511107 476021 476133 476806

0.979 0.993 0.930 0.964 1.219** 1.197** 1.134 1.109 0.968 0.718*

1.255 1.874*** 2.351*** 2.170***

1.479*

1.046 0.892 0.798**

0.979

1.021 0.919

Transition to parttime employment

1208113 810100 763052 763164 763837

1.075 1.382*** 2.052*** 2.356*** 2.816*** 2.843*** 2.671*** 2.399*** 2.006*** 1.340**

1.222 1.118 1.376*** 0.995

*

1.309*

1.045 1.043 1.016

0.931

1.022 1.077

Transition to marginal employment

1208113 7414805 7158502 7158614 7159287

1.144*** 1.297*** 1.355*** 1.164** 0.867* 0.781*** 0.855 0.828 0.737** 0.965

2.194*** 1.041 1.290*** 0.990

1.144

0.926 0.827** 0.718***

0.995

1.036 0.939

Unemployment

Analyses are based on all employment periods of persons who took up employment between 1.1.2000 and 31.12.2010. The period of observation was from 1.1.2000 to 31.12.2010. *p < 0.05; **p < 0.01; ***p < 0.001. Hazard ratios for the Cox partial-likelihood estimates; Calculation of cluster-robust standard errors for 96 planning regions. Source: LIAB (LM 9310), own calculations.

1208113 1385110 1369319 1369431 1370103

0.945 0.818** 0.950 0.972 1.002 0.876 0.839* 0.740*** 0.551*** 0.328***

1.511 1.714*** 1.894*** 1.413***

***

0.943 0.821* 0.709***

1.072** 1.035 0.972

***

0.923

0.902

0.962 1.070 0.973 0.814

0.980 0.830

not bound by a collective agreement) 1.006 0.970 0.981 0.942

Downward mobility

Collective agreement (Ref.: establishment Sectoral collective agreement Company collective agreement Orientation of sectoral collective agreement Firm size (Ref.: small firm) Small-medium-sized firm Medium-sized firm Larger firm Sector (Ref.: manufacturing industry) Agriculture, forestry, and mining Construction Trade Services for firms Other services

Lateral mobility

Exit from job

Independent variables

Table 22 (continued)

M. Dütsch, et al.

Advances in Life Course Research 40 (2019) 43–84

1.003 1.002 1.008 1.023 0.925 2.158 share of low-skilled workers) 1.007 0.993 1.011 0.948

Human capital endowment (Ref.: share of low-skilled workers) Share of skilled workers Share of highly skilled workers Share of unemployment periods Share of unemployment periods × human capital endowment (Ref.: Share of unemployment periods × share of skilled workers Share of unemployment periods × share of highly skilled workers 1.008 1.041** 1.217 0.975* 0.961**

0.961** 0.965

0.939*** 0.957*

0.990 1.028*** 1.131

0.990 0.991

1.029** 0.997

1.000 0.994 1.008 1.013 1.553

0.982* 1.023** 0.453

82

0.046*** 0.726 0.006*** 0.002*** 0.078* 1.034*** 0.981

1.067*** 1.027 1.040** 1.060

0.019*** 5.362*** 0.347 0.001*** 0.006*** 1.051*** 0.960** 1.012 1.080*** 1.060*** 1.032** 0.940* 0.993 1.016 1.027

0.335 1.879 0.463 0.007*** 0.227 1.017 0.970 1.009 1.059*** 1.031 0.984 1.025 1.028 1.080** 0.974

Analyses are based on all employment periods of persons who took up employment between 1.1.2000 and 31.12.2010. The period of observation was from 1.1.2000 to 31.12.2010. Models additionally contain all explanatory variables that are reported in Table 3. *p < 0.05; **p < 0.01; ***p < 0.001. Hazard ratios for the Cox partial-likelihood estimates; Calculation of cluster-robust standard errors for 96 planning regions. Source: LIAB (LM 9310), own calculations.

1.014 0.832**

1.055**

1.052***

0.995 0.995

0.920*** 0.968*

1.042*** 1.029* 5.566***

0.969*** 0.986

1.023* 1.025* 3.056***

0.994 1.009

1.024* 1.020 1.368

Unemployment

0.984 1.001

1.020 1.028

1.018 1.016 0.407

0.971 1.009

1.021 1.016 1.353

1.022** 0.995

1.008 1.021* 0.318

Transition to marginal employment

0.989 1.041***

0.999 0.998

1.004 1.031 1.900

1.012 1.037*

1.001 1.030** 0.341

0.993 0.971

1.042*** 1.017

1.003 1.019 0.216**

1.007 1.054*** 5.972***

Transition to part-time employment

0.959*** 1.014 0.019***

Transition to full-time employment

1.001 1.018 0.944

0.994 1.002

1.007 1.024 1.945

Upward mobility

Human capital endowment (Ref.: share of low-skilled workers) 1.001 0.967** 0.989 0.942*** Share of skilled workers 0.988* Share of highly skilled workers 0.998 0.998 0.998 1.002 1.006 Highest education level (Ref.: at most secondary school and no vocational training) At most secondary school and vocational training 0.099*** 0.493 0.226* 0.209 0.013*** Advanced secondary school and no vocational training 4.860 0.231 0.121 0.246 0.223 1.002 0.068** 0.354 0.007*** Advanced secondary school and vocational training 0.270* Polytechnic degree 0.196*** 4.822 0.131* 0.310 0.006*** University degree 0.562 0.855 0.064* 0.067** 0.612 Highest education level (Ref.: at most secondary school and no vocational training) × human capital endowment (Ref.: share of low-skilled workers) 1.003 1.011 1.017 1.056*** At most secondary school and vocational training × share of 1.026*** skilled workers Advanced secondary school and no vocational training × share 0.990 1.010 1.020 1.016 1.018 of skilled workers * Advanced secondary school and vocational training × share of 1.013 0.989 1.032 1.012 1.065*** skilled workers Polytechnic degree × share of skilled workers 1.014* 0.973 1.052** 0.989 1.052** University degree × share of skilled workers 1.009 1.002 1.062*** 1.057** 1.029* At most secondary school and vocational training × share of 1.019** 1.023 1.014 1.037 1.021 highly skilled workers Advanced secondary school and no vocational training × share 1.007 1.007 1.099* 1.018 1.028 of highly skilled workers * * ** ** Advanced secondary school and vocational training × share of 1.023 1.051 1.055 1.067 1.026 highly skilled workers Polytechnic degree × share of highly skilled workers 1.023* 1.011 1.002 1.000 1.080* 0.983 0.993 1.016 0.983 University degree × share of highly skilled workers 0.891**

1.003 1.002 1.011* 1.018 0.517 0.802 of low-skilled workers) 1.013 0.988 0.981 0.976

0.989 0.995

0.976 0.993

Human capital endowment (Ref.: share of low-skilled workers) Share of skilled workers Share of highly skilled workers Nationality (1 = foreign) Nationality (1 = foreign) × human capital endowment (Ref.: share Foreign × share of skilled workers Foreign × share of highly skilled workers

1.005 1.019 2.330

Downward mobility

1.011 1.019 6.387**

Exit from job Lateral mobility

Human capital endowment (Ref.: share of low-skilled workers) Share of skilled workers 1.009 Share of highly skilled workers 1.008 Sex (1 = female) 2.054* Sex (1 = female) × human capital endowment (Ref.: share of low-skilled workers) Female × share of skilled workers 0.989* Female × share of highly skilled workers 1.004

Independent variables

Table 23 Estimates for cross-level interactions between regional human capital endowment and individual factors influencing employment trajectories using imputed right censored wages.

M. Dütsch, et al.

Advances in Life Course Research 40 (2019) 43–84

Advances in Life Course Research 40 (2019) 43–84

M. Dütsch, et al.

References

job stability in Germany. European Societies, 6(1), 49–70. Esser, H. (1996). Soziologie: Allgemeine Grundlagen. Frankfurt: Campus. Evertsson, M., & Grunow, D. (2012). Women's work interruptions and career prospects in Germany and Sweden? International Journal of Sociology and Social Policy, 32(9), 561–575. Fitzenberger, B., Osikominu, A., & Völter, R. (2006). Imputation rules to improve the education variable in the IAB employment subsample. Schmollers Jahrbuch: Journal of Applied Social Science Studies, 126(3), 405–436. Fujita, M., Krugman, P., & Venables, A. (2001). The spatial economy: Cities, regions, and international trade. Cambridge: MIT Press. Gartner, H. (2005). The imputation of wages above the contribution limit with the German IAB employment sample. FDZ Methodenreport Nr. 2Nürnberg: Forschungsdatenzentrum (FDZ) der Bundesagentur für Arbeit. Gerlach, K., Meyer, W., & Tsertsvadze, G. (2002). Entwicklung der qualifikatorischen Segregation im Verarbeitenden Gewerbe. In L. Bellmann, & A. Kölling (Eds.). Betrieblicher Wandel und Fachkräftebedarf (pp. 51–84). Nürnberg: Institut für Arbeitsmarkt- und Berufsforschung. Giannelli, G., Jaenichen, U., & Rothe, Th. (2013). Doing well in reforming the labour market? Recent trends in job stability and wages in Germany. IZA Discussion Paper 7580. Giesecke, J., & Heisig, J. P. (2011). Destabilization and destandardization: For whom? The development of West German job mobility since. Schmollers Jahrbuch: Journal of Applied Social Science Studies, 131(2), 301–314. Glaeser, E. (1999). Learning in cities. Journal of Urban Economics, 46(2), 254–277. Glaeser, E., & Maré, D. (2001). Cities and skills. Journal of Labor Economics, 19(2), 316–342. Glaeser, E., & Saiz, A. (2004). The rise of the skilled city. Brookings-Wharton Papers on Urban Affairs. 47–105. Grotheer, M., Struck, O., Bellmann, L., & Gewiese, T. (2004). Determinanten der Beschäftigungsstabilität. Chancen und Risiken von “Entrants” im Ost-West-Vergleich. In O. Struck, & C. Köhler (Eds.). Beschäftigungsstabilität im Wandel? Empirische Befunde und theoretische Erklärungen für Ost- und Westdeutschland (pp. 181–198). Mering: Rainer Hampp. Gruhl, A., Schmucker, A., & Seth, S. (2012). The establishment history panel 1975–2010. Handbook version 2.2.1. FDZ-Datenreport, 04/2012 Nürnberg. Grunow, D., Schulz, F., & Blossfeld, H.-P. (2012). What determines change in the division of housework over the course of marriage? International Sociology, 27(3), 289–307. Harris, J., & Todaro, M. (1970). Migration, unemployment & development: A two-sector analysis. American Economic Review, 60(1), 126–142. Heinz, W., Huinink, J., & Weymann, A. (2009). The life course reader: Individuals and societies across time. Frankfurt am Main: Campus. Hendry, Ch. (2003). Applying employment systems theory to the analysis of national models of HRM. The International Journal of Human Resource Management, 14(8), 1430–1442. Hirsch, B., König, M., & Möller, J. (2013). Is there a gap in the gap? Regional differences in the gender pay gap. Scottish Journal of Political Economy, 60(4), 412–439. Hirschman, A. (1958). Strategy of economic development. New Haven: Yale University Press. Jovanovic, B. (1979). Job matching and the theory of turnover. Journal of Political Economy, 87(5), 972–990. Jovanovic, B., & Rob, R. (1989). The growth and diffusion of knowledge. The Review of Economic Studies, 56(4), 569–582. Kalbfleisch, J., & Prentice, R. (2002). The statistical analysis of failure time data. New Jersey: Wiley. Kalleberg, A., & Sørensen, A. (1979). The sociology of labor markets. Annual Review of Sociology, 5, 351–379. Kelle, M. (2016). Wage growth, urbanization, and firm characteristics – Evidence for Germany. Ruhr Economic Papers 631. Kézdi, G. (2004). Robust standard error estimation in fixed-effects panel models. Hungarian Statistical Review, 89, 96–116. Klosterhuber, W., Heining, J., & Seth, S. (2013). Linked-Employer–Employee data from the IAB: LIAB longitudinal model 1993–2010 (LIAB LM 9310). FDZ-Datenreport 08. Kogan, I. (2004). Labour market careers of immigrants in Germany and Great Britain. Journal of International Migration and Integration, 5(4), 419–449. Kogan, I. (2011). The price of being an outsider: Labour market flexibility and immigrants’ employment paths in Germany. International Journal of Comparative Sociology, 52(4), 264–283. Kohli, M. (1985). Die Institutionalisierung des Lebenslaufs. Kölner Zeitschrift für Soziologie und Sozialpsychologie, 37(1), 1–29. Konietzka, D. (2010). Berufliche Ausbildung und der Übergang in den Arbeitsmarkt. In R. Becker, & W. Lauterbach (Eds.). Bildung als Privileg: Erklärungen und Befunde zu den Ursachen der Bildungsungleichheit (pp. 277–304). Wiesbaden: VS Verlag für Sozialwissenschaften. Krugman, P. (1991). Geography and trade. Cambridge: MIT Press. Lepak, D., Hui, L., Yunhyung, Ch., & Harden, E. (2006). A conceptual review of human resource management systems in strategic human resource management research. Research in Personnel and Human Resources Management, 25, 217–271. Levy, R., & Bühlmann, F. (2016). Towards a socio-structural framework for life course analysis. Advances in Life Course Research, 30, 30–42. Lucas, R. (1988). On the mechanics of economic development. Journal of Monetary Economics, 22(1), 3–42. Manzoni, A. (2012). In and out of employment: Effects in panel and life-history data. Advances in Life Course Research, 17, 11–24. Manzoni, A., & Mooi-Reci, I. (2011). Early unemployment and subsequent career complexity: A sequence-based perspective. Schmollers Jahrbuch, 131(2), 339–348. Matano, A., & Naticchioni, P. (2016). What drives the urban wage premium? Evidence along the wage distribution. Journal of Regional Science, 56(2), 191–209.

Acemoglu, D. (1996). A microfoundation for social increasing returns in human capital accumulation. The Quarterly Journal of Economics, 111(3), 779–804. Acemoglu, D. (1999). Changes in unemployment and wage inequality: An alternative theory and some evidence. American Economic Review, 89(5), 1259–1278. Acemoglu, D. (2002). Technical change, inequality, and the labor market. Journal of Economic Literature, 40(1), 7–72. Andersson, M., Klaesson, J., & Larsson, J. P. (2014). The sources of the urban wage premium by worker skills: Spatial sorting or agglomeration economies? Papers in Regional Science, 93(4), 727–747. Angrist, J. D., & Pischke, J. S. (2009). Mostly harmless econometrics: An empiricist's companion. Princeton: Princeton University Press. Arrow, K. (1971). The theory of discrimination (Working Paper Industrial Relations Section, 30A). Princeton University. Arulampalam, W. (2001). Is unemployment really scarring? Effects of unemployment experiences on wages. Economic Journal, 111(475), 585–606. Baron, J., & Bielby, W. (1980). Bringing the firms back in: Stratification, segmentation, and the organization of work. American Sociological Review, 45(5), 737–765. BBSR (2018). INKAR 2018. Erläuterungen zu den Raumbezügen. Bonn: Bundesinstitut für Bau-, Stadt- und Raumforschung im Bundesamt für Bauwesen und Raumordnung http://www.inkar.de/documents/Erlaeuterungen%20Raumbezuege.pdf. Becker, G. (1962). Investment in human capital: A theoretical analysis. Journal of Political Economy, 70(5), 9–42. Becker, G. (1975). Human capital: A theoretical and empirical analysis, with special reference to education. New York: Columbia University Press. Berg, I. (1981). Sociological perspectives on labor markets. Quantitative studies in social relations. New York: Academic Press. Bergemann, A., & Mertens, A. (2004). Job stability trends, layoffs, and transitions to unemployment: An empirical analysis for West Germany. IZA-Diskussion Paper 1368. Blien, U., Kirchhof, K., & Ludewig, O. (2006). Agglomeration effects on labour demand. IABDiscussion Paper 28. Blien, U., & Wolf, K. (2002). Regional development of employment in eastern Germany. An analysis with an econometric analogue to shift-share techniques. Papers in Regional Science, 81(3), 391–414. Blossfeld, H.-P. (1985). Berufseintritt und Berufsverlauf: Eine Kohortenanalyse über die Bedeutung des ersten Berufs in der Erwerbsbiographie. Mitteilungen aus der Arbeitsmarkt- und Berufsforschung, 18(2), 177–197. Blossfeld, H.-P., Mills, M., & Bernardi, F. (2006). Globalization, uncertainty and men's careers: An international comparison. Cheltenham: Elgar. Blossfeld, H.-P., & Rohwer, G. (2002). Techniques of event history modelling: New approaches to causal analysis. Mahwah: Lawrence Erlbaum. Boockmann, B., & Steffes, S. (2005). Individual and plant-level determinants of job durations in Germany. ZEW-Discussion Paper 89. Boockmann, B., & Steffes, S. (2010). Workers, firms or institutions: What determines job duration for male employees in Germany? Industrial and Labor Relations Review, 64(1), 109–127. Bresnahan, T., Brynjolfsson, E., & Hitt, L. (2002). Information technology, workplace organization, and the demand for skilled labor: Firm-level evidence. The Quarterly Journal of Economics, 117(1), 339–376. Buch, T., Hamann, S., Niebuhr, A., & Rossen, A. (2017). How to woo the smart ones? Evaluating the determinants that particularly attract highly qualified people to cities. Journal of Urban Affairs, 39(6), 764–782. Buchholz, S. (2008). Die Flexibilisierung des Erwerbsverlaufs. Eine Analyse von Einstiegs- und Ausstiegsprozessen in Ost- und Westdeutschland (Dissertation Universität Bamberg) Wiesbaden: VS Verlag. Coleman, J. S. (1990). Foundations of social theory. Cambridge: Belknap Press of Harvard University Press. Combes, P.-P., Duranton, D., & Gobillon, L. (2008). Spatial wage disparities: Sorting matters!. Journal of Urban Economics, 63(2), 723–742. Combes, P.-P., & Gobillon, L. (2015). The empirics of agglomeration economies. In J. V. Henderson, & J. F. Thisse (Eds.). Handbook of regional and urban economics, Vol. 4. Cities and geography (pp. 247–348). Amsterdam U.A: Elsevier. Cox, D. (1972). Regression models and life-tables. Journal of the Royal Statistical Society, 34, 187–202. Dauth, W. (2013). Agglomeration and regional employment growth. Papers in Regional Science, 92(2), 419–435. D’Costa, S., & Overman, H. G. (2014). The urban wage growth premium: Sorting or learning? Regional Science and Urban Economics, 48, 168–179. De La Roca, J., & Puga, D. (2017). Learning by working in big cities. The Review of Economic Studies, 84(1), 106–142. Diewald, M., & Sill, S. (2014). Mehr Risiken, mehr Chancen? Trends in der Arbeitsmarktmobilität seit Mitte der 1980er Jahre. In O. Struck, & Ch. Köhler (Eds.). Beschäftigungsstabilität im Wandel? (pp. 39–62). München and Mering: Hampp. Duranton, G. (2004). The economics of production systems: Segmentation and skillbiased change. European Economic Review, 48(2), 307–336. Duranton, G., & Puga, D. (2004). Micro-foundations of urban agglomeration economies. In J. V. Henderson, & J. F. Thisse (Eds.). Handbook of regional and urban economics, Vol. 4. Cities and geography (pp. 2063–2117). Amsterdam U.A: Elsevier. Dütsch, M., & Struck, O. (2014). Employment trajectories in Germany. Do firm characteristics and regional disparities matter? Journal for Labour Market Research, 47(1/ 2), 107–127. Erlinghagen, M. (2004). Die Restrukturierung des Arbeitsmarktes: Arbeitsmarktmobilität und Beschäftigungsstabilität im Zeitverlauf. Wiesbaden: VS Verlag. Erlinghagen, M., & Knuth, M. (2004). In search of turbulence. Labour market mobility and

83

Advances in Life Course Research 40 (2019) 43–84

M. Dütsch, et al. Mayer, K. U. (2009). New directions in life course research. Annual Review of Sociology, 35, 413–433. Melo, P. C., & Graham, D. J. (2014). Testing for labour pooling as a source of agglomeration economies: Evidence for labour markets in England and Wales. Papers in Regional Science, 93(1), 31–52. Mincer, J., & Polachek, S. (1974). Family investments in human capital: Earnings of women. The Journal of Political Economy, 82(2), 76–108. Mumford, K., & Smith, P. (2004). Job Tenure in Britain: Employee characteristics versus workplace effects. Economica, 71(282), 275–298. Phelps, E. (1972). The statistical theory of racism and sexism. American Economic Review, 62(4), 659–661. Poelhekke, S. (2013). Human capital and employment growth in German metropolitan areas: New evidence. Regional Studies, 47(2), 245–263. Rendtel, U., & Schwarze, J. (1996). Schätzungen von Lohnkurven für Westdeutschland mit einem verallgemeinerten Varianz-Komponenten-Modell. Mitteilungen aus der Arbeitsmarkt- und Berufsforschung, 3, 491–493. Romer, P. (1990). Endogenous technological change. Journal of Political Economy, 98(5), 71–102. Schlitte, F. (2012). Local human capital, segregation by skill, and skill-specific employment growth. Papers in Regional Science, 91(1), 85–106. Schwarze, J. (1995). Neue Befunde zur “Lohnkurve” in Deutschland: Eine Analyse mit Paneldaten für Raumordnungsregionen 1985 und 1989. DIW Diskussionspapier 119. Shapiro, J. (2006). Smart cities: Quality of life, productivity, and the growth effects of human capital. The Review of Economics and Statistics, 88(2), 324–335. Simon, C., & Nardinelli, C. (2002). Human capital and the rise of American cities, 1900–1990. Regional Science and Urban Economics, 32(1), 59–96. Spence, M. (1973). Job marketing signaling. Quarterly Journal of Economics, 87(3), 355–374. Spence, M. (2002). Signaling in retrospect and the informational structure of markets. American Economic Review, 92(3), 434–459.

Stephan, G. (2001). Firmenlohndifferenziale: Eine Analyse für die Bundesrepublik Deutschland. Studien zur Arbeitsmarktforschung. Bd. 13. Frankfurt am Main: Campus. Struck, O., Grotheer, M., Schröder, T., & Köhler, Ch. (2007). Instabile Beschäftigung: Neue Ergebnisse zu einer alten Kontroverse. Kölner Zeitschrift für Soziologie und Sozialpsychologie, 59(2), 294–317. Südekum, J. (2008). Convergence of the skill composition across German regions. Regional Science and Urban Economics, 38(2), 148–159. Treas, J., & Lui, J. (2013). Studying housework across nations. Journal of Family Theory & Review, 5(2), 135–149. Wheeler, C. (2001). Search, sorting, and urban agglomeration. Journal of Labor Economics, 19(4), 879–899. Wheeler, C. (2006). Cities and the growth of wages among young workers: Evidence from the NLSY. Journal of Urban Economics, 60, 162–184. Wheeler, C. (2008). Local market scale and the pattern of job changes among young men. Regional Science and Urban Economics, 38, 101–118. Winkelmann, R., & Zimmermann, K. (1998). Is job stability declining in Germany? Evidence from count data models. Applied Economics, 30(11), 1413–1420.

Further reading Mincer, J. (1962). On-the-job training: Costs, returns and some implications. Journal of Political Economy, 70, 50-79. Farhauer, O., Granato, N. (2006). Regionale Arbeitsmärkte in Westdeutschland: Standortfaktoren und Branchenmix entscheidend für Beschäftigung, IAB Kurzbericht, 4. Krugman, P. (1998). What’s new about the new economic geography? Oxford Review of Economic Policy, 14 (2), 7-17.

84