Youth unemployment and employment trajectories in Spain during the Great Recession: what are the determinants?

Verd, Joan Miquel; Barranco, Oriol; Bolíbar, Mireia Article Youth unemployment and employment trajectories in Spain during the Great Recession: what are the determinants? Journal for Labour Market Research Provided in Cooperation with: Institute for Employment Research (IAB) Suggested Citation: Verd, Joan Miquel; Barranco, Oriol; Bolíbar, Mireia (2019) : Youth unemployment and employment trajectories in Spain during the Great Recession: what are the determinants?, Journal for Labour Market Research, ISSN 2510-5027, Springer, Heidelberg, Vol. 53, Iss. 4, pp. 1-20, https://doi.org/10.1186/s12651-019-0254-3 This Version is available at: https://hdl.handle.net/10419/216678 Standard-Nutzungsbedingungen: Die Dokumente auf EconStor dürfen zu eigenen wissenschaftlichen Zwecken und zum Privatgebrauch gespeichert und kopiert werden. Sie dürfen die Dokumente nicht für öffentliche oder kommerzielle Zwecke vervielfältigen, öffentlich ausstellen, öffentlich zugänglich machen, vertreiben oder anderweitig nutzen. Sofern die Verfasser die Dokumente unter Open-Content-Lizenzen (insbesondere CC-Lizenzen) zur Verfügung gestellt haben sollten, gelten abweichend von diesen Nutzungsbedingungen die in der dort genannten Lizenz gewährten Nutzungsrechte. Terms of use: Documents in EconStor may be saved and copied for your personal and scholarly purposes. You are not to copy documents for public or commercial purposes, to exhibit the documents publicly, to make them publicly available on the internet, or to distribute or otherwise use the documents in public. If the documents have been made available under an Open Content Licence (especially Creative Commons Licences), you may exercise further usage rights as specified in the indicated licence. https://creativecommons.org/licenses/by/4.0/ Verdetal. J Labour Market Res (2019) 53:4 https://doi.org/10.1186/s12651-019-0254-3 ARTICLE Youth unemployment andemployment trajectories inSpain duringtheGreat Recession: what are thedeterminants? Joan Miquel Verd1* , Oriol Barranco1 and Mireia Bolíbar2 Abstract Since the beginning of the recession period in Europe, unemployment has greatly affected the young adult population. In this context, Spain is regarded as an extreme case, due to its exceptionally high youth unemployment rates. This article seeks to identify the determinants that have led certain groups of Spanish young people to suffer labour market trajectories with higher levels of unemployment and instability during the Great Recession than others. To do this, retrospective data from the 2012 Catalan Youth Survey are used. With these data and using cluster analysis, a typology of labour market trajectories is constructed. Next, multinomial logistic regressions are used to identify what individual socio-demographic characteristics and pre-crisis employment experiences are connected to these different typological career paths. Results show that the highly differentiated career paths are associated with different social profiles and differences in the presence of unemployment. Moreover, interesting differences among the most unstable career paths appear. For the most vulnerable social profiles the employment trajectory prior to the crisis seems to point towards the existence of an entrapment in low-skilled jobs that alternate with situations of unemployment. For those with a slightly better position their employment situation after the initiation of the crisis seems to have been impacted by their brief labour market trajectory before the crisis and their resulting work experience gap. Keywords: Spain, Unemployment, Great Recession, Youth, Labour markets, Employment trajectories JEL Classification: J24, J64, Z13 © The Author(s) 2019. This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creat iveco mmons .org/licen ses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. 1 Introduction Spain, along with Greece, Italy and Croatia, is one of the countries in the European Union where unemployment has hit young people the hardest during the Great Recession. In its aftermath, unemployment among Spanish workers under 25years of age reached peaks above 50% in 2012, 2013 and 2014. Since these peaks, the youth unemployment rate has been slowly declining, although in 2017 it continued to be above 35% (Spanish Labour Force Survey). The destruction of youth employment in Spain during the Great Recession must be explained by the sensitivity of this population group to economic downturns, which historically and all over western countries provoke a growth in unemployment above the average found for the overall population (Lefresne 2003; Verick 2011; Choudhry et al. 2012; ILO 2016). However, the existence of country-specific characteristics that produce enormous variations among the countries of the European Union must also be taken into consideration (Pastore and Giuliani 2015; Dietrich and Möller 2016). In addition, although in Spain many have come to speak of a “lost generation”—more so in the media than in the strictly academic world—referring to all young people in a generic manner, there are major differences in the rates of unemployment among young people with different social profiles. As a result, to obtain a complete picture Open Access Journal for Labour Market Research *Correspondence: joanmiquel.v[email protected] 1 Centre d’Estudis Sociologics Sobre la Vida Quotidiana i el Treball (Sociological Research Centre on Everyday Life and Work-QUIT), Institut d’Estudis del Treball (Institute for Labour Studies-IET), Universitat Autonoma de Barcelona, Building B. Campus of the Universitat Autonoma of Barcelona, 08193 Cerdanyola del Vallés, Barcelona, Spain Full list of author information is available at the end of the article Page 2 of 20 Verdetal. J Labour Market Res (2019) 53:4 of youth unemployment it is necessary to also take into account the socio-demographic characteristics of young people and their type of insertion in the labour market (Freeman and Wise1982). The primary objective of this article is to identify, from a sociological perspective, what determinants are associated with youth unemployment in Spain during the Great Recession, addressing both socio-demographic characteristics as well as factors connected to career paths. Moreover, the article has the objective of framing the analysis of unemployment in the kind of labour market trajectory developed by young workers. Thus, instead of considering unemployment as a single dependent variable, the article considers periods of unemployment in connection to other events in the individual’s labour market trajectory, making explicit in this way that unemployment should not be understood as an event isolated from the labour market segment in which career paths develop. Accordingly, the analysis presented is designed to discern between the influence of what we call ‘social factors’ (socio-demographic characteristics of individuals) and ‘career factors’ (the length and quality of the labour market experience) on the kind of labour market trajectory undertaken during the Great Recession. To fully develop these initial objectives the article adopts a dynamic perspective. Firstly, employment trajectories are constructed by means of a cluster analysis to permit us to identify different types of career paths with different levels of unemployment and instability associated with them. Secondly, multinomial logistic regressions are used to examine the degree of association of these different trajectories with different sub-groups of the working youth population and the features of their trajectories in the labour market during the period previous to recession. This approach makes it possible to highlight certain new determinants, beyond individual characteristics, of the hardships faced by young people during the Great Recession in Spain. Our starting hypothesis is that during the crisis, unemployment and unstable employment have most severely affected the trajectories of young people with sociodemographic profiles that have traditionally been more vulnerable in the Spanish labour market, as well as those who already had experienced unstable and low quality employment before the crisis. By adopting an approach based on an analysis of employment trajectories, this article improves on existing studies on the characteristics of youth unemployment in Spain. One of the relevant findings is that long-term and short-term unemployment during the Great Recession correspond to different kinds of labour market trajectory. In addition, we find that a long employment experience prior to the onset of recession is associated with having a stable career path during recession, thus having a protective effect against unemployment. However, this is the case only for those workers not employed in the most precarious labour market segment. For the most vulnerable social profiles the length of the employment trajectory prior to the crisis had no effect. Therefore, the research reported in the article shows that the determinants of youth unemployment are better understood when taking into account the existence of different layers of labour market segmentation in Spain, and not only the difference between primary and secondary segmentsor stable and unstable segments. To our knowledge, there is no published research yet evaluating the role of unemployment in the development of employment trajectories among young people during the Great Recession in Spain. Our analysis complements that of López-Andreu and Verd (2016), who use panel data1 to analyse the segmentation of career paths among young adult workers (individuals between 20 and 40years old) for the period between 2007 and 2011. Their analysis, which uses a multinomial logit model, identifies the importance of industrial sectors and family background in the development of trajectories for the population considered. For the period prior to the recession the few existing analyses that have addressed the question of employment trajectories (Hernanz 2003; García-Pérez and Muñoz-Bullón 2011; Verd and López-Andreu 2012)2 have revealed the important role that the combination of unemployment and temporary contracts plays in the trajectories of employees with low educational levels, trapping them in unskilled jobs from which it is difficult to escape. The novelty of our analysis is that in addition to this most precarious trajectory—which continues to exist during the Great Recession—we find that unemployment has also hit less vulnerable profiles, whose situation of unemployment cannot be explained by only their entrapment in low-skilled jobs, but also by their lack of sufficient experience in the labour market. The identification of this intermediate segment (not stable, but not completely precarious), where the associated social profiles suffered unemployment to a lesser degree than vulnerable profiles, can be considered a contribution to the 1 The panel survey used as a source of data in their article (the Inequalities Panel of the Jaume Bofill Foundation) was discontinued in 2012. Although the data we use is not panel data, it has the advantage of providing month to month information, while the latter one provided only annual information (for instance, in terms of employment, the information provided was the most frequent situation along every year considered). 2 Hernanz (2003) uses the linked Spanish Labour Force Survey (1987– 2001) for her analysis, García-Pérez and Muñoz-Bullón (2011) use the Spanish Social Security records for the period 1996–2003, and Verd and López-Andreu (2012) use the Inequalities Panel of the Jaume Bofill Foundation—already mentioned in note 1—for the period 2001–2006. Page 3 of 20 Verdetal. J Labour Market Res (2019) 53:4 existing literature, revealing a new feature in relation to how unemployment impacted workers in previous economic crises. Using a dynamic perspective poses an important methodological difficulty in the Spanish case: the lack of longitudinal databases that gather adequate data on situations of unemployment and family social background as well. To overcome this problem, we use the retrospective data provided by the 2012 Catalan Youth Survey (Enquesta a la joventut de Catalunya 2012). This survey provides a long list of relevant economic and social variables to study different dimensions of the lives of a sample of 3002 respondents. This is not a panel survey, but permits us to reconstruct the employment trajectories of young people since they were 15years of age by means of retrospective questions; in addition it provides information on their family background. Catalonia accounted for 19.9% of Spain’s GDP and 17.4% of the country’s total employment in 2007. For the year 2017 these figures where 19.2% for GDP and 17.4% for the share of Spanish employment. Catalonia is also one of the regions with the lowest unemployment and temporary employment rates in Spain. However, these differences with the Spanish labour market as a whole are mostly quantitative—in intensity and magnitude—but not qualitative—as the key dynamics and determinants affecting the level of unemployment and temporary employment are shared by all Spanish regions. In Sect.3.1 we provide grounds for this argument. Thus, in terms of representativeness, the relative frequencies of the different labour market trajectories analysed in the article should not be extrapolated to the whole Spanish level, however, we do believe that the conclusions regarding the determinants linked to each type of trajectory can, to a great extent, be extrapolated to the whole Spain. This should not be surprising, as the regulatory and institutional framework is the same for all regions, social stratification is similar, and the impact of the recession on the evolution of youth unemployment levels has also been very similar (see Table1). The rest of this article is structured in the following manner. In Sect.2 we present a brief theoretical review of the literature that has addressed the main determinants of youth unemployment generally as well as for the Spanish case. In Sect.3 we present our data and the methods of analysis used. In Sect.4 we present the results of our analysis and our interpretation of them. Lastly, in Sect. 5, the concluding section, we briefly summarise these findings, and reflect upon the prospects these results may imply for Spanish youth and for policy making. 2 Unemployment inthetrajectories oftheworking youth population inSpain: social factors andcareer factors Table1 shows the evolution of unemployment for Catalonia and Spain from 2006 to 2013. As can be seen, although the rates are somewhat lower in Catalonia, the impact of the Great Recession in both contexts follows an identical trend, and in both cases, the increase in youth unemployment (young people from 16 to 24years of age) is spectacular. In 2006, youth unemployment was 17.9% for Spain overall and 14.7% in Catalonia. In 2008, with the crisis having just begun, youth unemployment was 24.6% and 20.4% respectively, and in 2013, the height of youth unemployment during the Great Recession, the rates were 54.9% and 50% respectively, rates that are more than triple that found in 2006, prior to the crisis. What factors could account for this enormous growth in unemployment among young people in Spain and Catalonia? There are many studies that have tried to explain, from a comparative perspective, the causes for the rise in youth unemployment to such levels during the economic crisis. On the one hand, some mention the widespread introduction of temporary contracts in the years prior to the crisis, as a result of reforms “at the margin” or “two-tier reforms”, which mostly affected young workers (Mertens etal. 2007; O’Reilly etal. 2015). Some authors consider these reforms to have consolidated a labour market marked by differences between insiders and outsiders (Bentolila etal. 2008), with the increase in unemployment during the crisis being caused by the large gap between the costs of firing persons with permanent versus temporary contracts and the laxer regulations on the use of the latter (Bentolila etal. 2012). From this perspective, it can be deduced that the higher rates of Table 1 Unemployment rates amongyoung people (16–24years ofage). Spain andCatalonia. 2006 to2013 Source: Elaborated by authors based on data from the Statistical Institute of Catalonia (IDESCAT) and the Spanish Labour Force Survey 2006 2007 2008 2009 2010 2011 2012 2013 Spain 17.9 18.2 24.6 37.8 41.6 46.4 53.2 54.9 Catalonia 14.7 13.5 20.4 37.1 39.5 44.1 50.7 50.8 Page 4 of 20 Verdetal. J Labour Market Res (2019) 53:4 unemployment suffered by young people are due to the greater concentration of temporary contracts among this group in comparison to prime-age workers.3 On the other hand, the particular harshness with which the recession has affected Spain has also been noted (O’Higgins 2012). Within this context of crisis businesses apply the principle of last-in first-out (LIFO) to reduce their workforce (Bell and Blanchflower 2011; Rocha 2012a; Dietrich and Möller 2016), which primarily affects young people. Lastly, the existence in Spain of a deficient institutional design—too rigid and sequential—in regard to schoolto-work transitions, has also been noted (Pastore 2015; Pastore and Giuliani 2015), which makes the acquisition of employment experience among young workers more difficult, making them less attractive to employers. Possibly, the real causes can be found in a combination of the factors just mentioned, some of them shared with other Mediterranean countries, that have also been hit with a strong increase in youth unemployment (Dietrich and Möller 2016; Pastore 2017). The above mentioned comparative approaches are useful for identifying country-specific factors that influence aggregate unemployment rates for young people in Spain. However, to obtain a complete picture of the unemployment problem it is important to connect the functioning of the labour market with the individual characteristics of young people. As stressed in the introduction, some young people are more likely than others to suffer a situation of unemployment. Thus, if we want to identify the causes for the concentration of unemployment in certain social profiles among young people, it is necessary to take into consideration the possible effect of the sociodemographic characteristics of this population and the influence of the type of insertion in the labour market, as Freeman and Wise (1982) argued some years ago. If we look at socio-demographic characteristics, analyses based on cross-sectional data have shown that the groups that have been most affected by the destruction of employment during the Great Recession are immigrants, women, those with low education level and workers under 25years of age (Rocha 2012a, b; Alós 2012; Alós and Lope 2015). It is not by chance that these groups are also those that occupy the most low-skilled segments of the Spanish labour market, in which unstable employment is the norm (Alós 2008, 2015). On the one hand, we must take into account that firms use temporary contracts to cover those positions in which increases in productivity are not associated with the permanence of the worker (Hernanz 2003). On the other hand, the spread of the norm of flexible employment has created an “imperfect dual market” (Prieto etal. 2009) that is extremely sensitive to economic cycles (Banyuls etal. 2009; Dietrich and Möller 2016; Muñoz-de-Bustillo and Esteve 2017). Thus, in a macroeconomic context such as in Spain during the recession, everything seems to indicate that firms have responded to the crisis by discontinuing or not renewing temporary contracts. Therefore, the weak position of certain social groups—in terms of the labour market segment where they are placed—could explain, in an initial and straightforward cross-sectional approach, why individuals employed at that point (mostly in temporary jobs) lost their jobs with the initiation of the crisis.4 However, a dynamic approach must be taken to obtain a better picture of the vulnerability of certain young workers in the Spanish labour market and whether they had a greater likelihood of becoming unemployed once the recession began. At the international level evidence exists that temporary employment is a stepping stone to permanent jobs for certain groups (Eichhorst etal. 2017), although this varies significantly among countries (Scherer 2004; Berton etal. 2011; Bruno etal. 2012). But temporary employment can also lead to dead end jobs, which result in becoming trapped in instability (Scherer 2004; Barbieri and Scherer 2009; Bruno et al. 2012) and, consequently, in having increased possibility of being dismissed in a context of a lack of employment.5 Thus, if we really want to identify, from a sociological perspective, the relationship between socio-demographic characteristics and the likelihood of becoming unemployed we need to take into account the employment trajectories of young people (i.e., to address their (un)employment situations from a longitudinal perspective). Once we know these trajectories, it will be possible to identify the factors that, in the context of the particular characteristics of the Spanish labour market, shape the type of career path developed by young people during the years of economic downturn in Spain. Moreover, a dynamic approach is necessary if we take fully into account the initial theoretical formulation of the theory of labour market segmentation (Doeringer and Piore 1971; Edwards 1975). The initial theory of segmentation stresses the temporal conception of segments, as in its understanding labour market segmentation implies that employment positions in the primary 4 Already in the 1980s Osterman (1980) and Kaliski (1984, p. 131) mentioned the preference that firms with better jobs had for contracting individuals with employment experience, which excluded young people entering the labour market. 5 Becoming trapped can also occur in relation to unskilled jobs. Baert etal. (2013) and Meroni and Vera-Toscano (2017) have shown that being overqualified at the beginning of a career can be a trap that increases subsequent difficulties in finding a job consistent with the individual’s skills. 3 However, the analyses of dualisation in different employment regimes in Europe carried out by Häusermann and Schwander (2012) do not place Spain among the countries with a strong insider–outsider model. Page 5 of 20 Verdetal. J Labour Market Res (2019) 53:4 segments are closed to those workers in a lower segment, so that the employment trajectories of workers develop entirely in one segment (Tomlinson and Walker 2012). Finally, the dynamic view will help us to identify the role played by career path in the years prior to the recession in relation to the employment trajectory developed after its onset. The trajectory of a person can be understood as a process that permits the accumulation of experience, skills and competencies valued by employers (Ryan 2001; Debels 2008; Pastore 2015). The question is if before a shock as significant as that produced by the economic crisis in Spain, this prior trajectory has an impact on firms in their decision to lay-off certain workers, or if lay-offs are carried out independently of previous career path. Clearly, individual trajectories prior to the crisis are not independent of individuals’ socio-demographic characteristics, making it necessary to also control for these factors. Examining these relations, it is possible to evaluate the degree of path dependency of the employment trajectory after the onset of the crisis. 3 Data andmethods 3.1 Data This article uses data from the Catalan Youth Survey 2012 (Enquesta a la joventut de Catalunya 2012). This survey is aimed at a representative sample of young people in Catalonia between 15 and 34years of age. It looks at their social characteristics, with particular emphasis on their trajectories in education, training and employment as well as processes of transition to adult life. It is produced by the Catalan Youth Observatory in collaboration with the Statistical Institute of Catalonia (IDESCAT). The survey sample size is 3002 individuals and the field work was carried out between the months of April and July of 2012. It was based on two-stage sampling by clusters with prior stratification. The units in the first stage are municipalities, stratified by territory and population size. In the second stage the units are the individuals, who are stratified by sex, age group (15–19, 20–24, 25–29 and 30–34years of age) and nationality (Spanish or foreign). The analysis for this article has been carried out on the sub-sample of young people that at the time of the beginning of the recession were or had been active in the labour market (N = 1394). In the analyses we have used data for individuals between 20 and 34years of age in order to avoid the distortions caused by the high percentage of young people under 20 still in education. Although the cut-off age of 34 to define youth may seem surprising in comparison with the age definition of youth adopted in the majority of other European countries, it is optimal for the analysis of Spanish youth (un)employment because it allows us to take into account the complexity of labour market insertion processes in Spain, which are prolonged over a greater period of time (Casal etal. 2006; Quintini etal. 2007, p.8). The survey provides relevant information for sociological analysis as it offers rich data on individuals’ social and family background. It includes variables on parents’ occupational category when respondents were 15years of age, in addition to standard variables like age, sex, place of birth and educational level. Table4 in Appendix contains the descriptive statistics for the variables used in the analyses. The survey also includes data on employment trajectories obtained from retrospective questions in the questionnaire. Specifically, the survey identifies the economic activities that compose respondents’ trajectories from when they were 15years of age until the point of the interview, specifying the moment they entered the labour market (month and year) and their current situation. This detailed retrospective information is especially useful for the analysis of employment trajectories. It has been used in the analyses to compute variables regarding (i) the quality of the work experience young individuals accumulated prior to the start of the recession, and, more importantly, (ii) the characteristics of their labour market trajectories during the period from October 2007 to April 2012. The latter are used to build a typology of labour market trajectories, while the former, together with socio-demographic variables, are used as covariates of the trajectory. Finally, it is important considering the degree of representativeness of the Catalan case to account for the determinants of unemployment and labour market trajectories of young individuals in Spain as a whole. In this sense, first, it must be pointed out that Spain has relevant regional disparities in unemployment and temporary employment rates, although unemployment disparities are less marked than in Germany or Italy (López-Bazo and Montellón 2013, p. 384). The acknowledgement of these regional disparities has led some authors to group the 17 Spanish regions into high and low unemployment regions (e.g. Bande and Karanassou 2014) as well as into two groups according to the presence of temporary employment (Caparrós and Navarro 2008).6 Taking into account both types of groupings, we can identify a 6 Spanish regional unemployment has been clustered in different studies (e.g. López-Bazo and Montellón 2013; Sala and Trivín 2014; Bande and Karanassou 2014). The clustering proposed by Bande and Karanassou (2014) is the one that uses data from the longest period: 1980–2010. These authors group Aragon, the Balearic Islands, Catalonia, Madrid, Navarre, the Basque Country and La Rioja into low unemployment regions, while they classify the remaining 10 Spanish regions as high unemployment regions. Caparrós and Navarro (2008), with data from 2000, group the former regions together as well along with Extremadura and Castilla-La Mancha, though leaving out the Basque Country, as the regions with a lower presence of temporary employment. Page 6 of 20 Verdetal. J Labour Market Res (2019) 53:4 cluster of regions with lower unemployment and temporary employment, which would be composed of Aragon, the Balearic Islands, Catalonia, Madrid, Navarre and La Rioja. The findings regarding the relative weight of the different labour market trajectories found in the analysis can only be extrapolated to these regions.7 Despite these differences in magnitude, the dynamics of unemployment and temporary employment seem to be the same in all the regions in Spain. In this regard, Sala and Trivín (2014) show that, for the 1996–2012 period, both low and high unemployment regions increased participation and employment levels (the latter was even slightly larger in high unemployment regions) during the period previous to economic crisis,while unemployment and spatial mobility rose up in a similar vein during the recession. Caparrós and Navarro (2008) show that in all Spanish regions being a woman, young, low-educated, low-skilled or employed in agriculture or construction sectors increase the probability of having temporary contracts. These data on the dynamics and factors linked to the labour market situation of individuals in the different Spanish regions lead us to conclude that our findings regarding the determinants linked to each type of trajectory can, to a great extent, be extrapolated to all of Spain. 3.2 Analysis The employment trajectories of the young people of the sample had differing lengths at the moment of the survey, depending on how long ago they first entered the labour market. However, the focus of the analysis is on their trajectories during the crisis period (from October 2007 to April 2012), in which all have the same length of 55months. Regarding the trajectories during the crisis, we consider the different situations of activity and employment over the course of this period as percentages of the individual’s trajectory during the crisis. Based on these data we have built a typology of labour market trajectories using cluster analysis. After, we have applied multinomial logistic regressions to describe the social and career factors associated with the type of labour trajectory young people have experienced during the crisis. The goal of the cluster analysis was to identify groups with similar labour markettrajectories. This was done by measuring the percentage of time (measured in months) spent in seven types of employment situations during the four years and seven months of the economic crisis taken into consideration in our analysis. The seven distinct situations considered in the analysis are unemployment, stable employment, temporary employment (includes jobs with fixed term contracts and paid internships), nonsalaried employment (includes all forms of employment outside of dependent employment, i.e. self employment with and without employees), short odd jobs (days, weeks or a few months), informal employment (without a contract, or working in a family business without pay), and inactive. In order to identify distinct groups based on similar trajectories as defined by these variables and to create a typology from them we have used the Ward’s hierarchical clustering method (Ward 1963) through use of SPSS software. This method consists of a hierarchical cluster algorithm in which the aggregation of cases seeks the minimum loss of variance in each step, based on an assessment of their similarity calculated as the sum of squares between every pair of clusters computed over all variables (Hair etal. 2009).8 The decision on the partitioning of the cluster in a solution of n groups, contrary to other’model-based clustering’procedures (such as Latent Profile Analysis) that draw on assumptions of data distribution, is not based on any confirmatory test. As Hair etal. (2009) explain, it is rather based on a combination of interpretive and technical-statistical criteria. In other words, it is aimed at selecting the simplest model possible that would capture the identity of each type within the social diversity the data reflect, while taking into consideration the analysis of the values of the clustering coefficients. The clustering process is shown visually in a dendrogram (see Fig.1 in Appendix), graphically depicting in the axis of abscissas the increase in heterogeneity at every step of the process of grouping cases and groups. In our case, of all possible solutions of 2, 3, 4, 5 or more clusters, we chose the four cluster solution because it allowed grouping internally homogeneous trajectories and showed a clearly defined profile for every group. The solution of 2 clusters, in spite of being the most optimal according to the clustering coefficients, merged trajectories that were too heterogeneous into a single group, 8 In the Ward’s procedure, the selection of which two clusters to combine at each step of the agglomerative process is based on which combination of clusters minimises the within-cluster sum of squares across the complete set of disjoint or separate clusters. At each step, the two combined clusters are those that minimise the increase in the total sum of squares across all variables in all clusters, therefore clusters with a small number of observations are usually merged. The Ward method is widely-used because of its tendency to generate clusters that are homogeneous and relatively equal in size (Hair etal. 2009, p. 505). 7 Another relevant difference—although not large—regarding our objectives is the sectoral distribution of employment. In 2016 in Catalonia this distribution was the following: 1.60% of employed persons working in agriculture, 18.33 working in industry, 5.79 working in building industry, and 74.29 in services. For the whole of Spain this distribution was 4.22% of employed persons working in agriculture, 13.75 working in industry, 5.85 working in building industry, and 76.17 working in services (data from the Spanish National Statistics Institute). Page 7 of 20 Verdetal. J Labour Market Res (2019) 53:4 thus blurring the interpretation of the results, which was clearer in the partition into 4 clusters.9 After, we ran two multinomial logistic regressions. The first contrasts the characteristics of young people with different types of trajectories and the factors that contributed to them, such as social and family origin. Specifically, the variables used in the first regression were sex, age, place of birth (Spain or elsewhere), educational level and parents’ occupational status. This variable was introduced using the dominance or dominant position criterion (Erikson 1984, p. 501; Korupp et al. 2002, p. 19), which establishes that the category which is highest internships, short term temporary work, non-contract work, or non-paid work in a family business) and never having experienced it (therefore having always had stable employment, defined by having had open-ended contracts or being self-employed). The absence of collinearity among predictors was checked. The multinomial regression models estimate the effect of socio-demographic and trajectory variables on the probability of experiencing the most insecure labour market trajectories against the probability of experiencing the most stable trajectory. The estimation equation is given as follows (where TPR stands for the trajectory prior to the recession): A robustness check has been performed by reproducing the regression analysis only with the sub-population of young people up to 29years of age. The effects of most of the predictors remain the same, except for place of birth and level of education, which do not exhibit statistical significance for explaining the trajectories when controlled by the variables of the nature of the pre-recession experience. Moreover, among this sub-population the length of the TPR is more important in the prediction of the non-salaried trajectory, while having experienced unemployment or unstable employment prior to the recession is not significant. 4 Results anddiscussion 4.1 Results As mentioned in the methods section, the four cluster solution we have retained allows grouping internally homogeneous trajectories that, at the same time, have very different characteristics. This can be observed in Table2, which presents a synopsis of the main characteristics of the four types of trajectories, including the most characteristic events of every type of trajectory, the kind of unemployment experienced in the trajectory (if any), and the socio-demographic characteristics of the logit(y=precarious trajectory)=log P(y = precarious trajectory) P(y=stable trajectory)  =α+βsex +βage +βplace of birth +βparents′occ.status +βlength TPR +βhighest occ.status TPR +βstability TPR logit(y=non−salaried trajectory)=log P(y=non−salaried trajectory) P(y=stable trajectory) =α+βsex +βage +βplace of birth +βparents′occ.status +βlength TPR +βhighest occ.status TPR +βstability TPR logit (y=temporary empl.trajectory)=log P(y=temporary empl.trajectory) P(y=stable trajectory) =α+βsex +βage +βplace of birth +βparents′occ.status +βlength PRT + βhighest occ.status TPR + βstability TPR 9 Hair etal. (2009, p. 524) warn against the selection of two-cluster solutions. Despite the fact that it often involves the largest increase in heterogeneity, a two-cluster solution often has limited heuristic value and should only be selected if supported by strong theoretical reasoning. 10 When speaking of our analysis, and also in Tables2, 3, 4, 7 and 8 we use the expression trajectory prior to the recession (TPR) to refer generally to these three variables. hierarchically (be it the mother or the father) should be chosen. The second regression connects the different types of trajectories with variables showing the length and quality of the labour experience young people had at the start of the recession.10 Specifically, the variables considered in addition to those used in the first model are labour market experience (i.e., years in the labour market at the start of the crisis (as a continuous variable), the highest occupational status acquired during the period prior to the start of the recession, and the experience of unemployment or unstable employment also in the trajectory prior to the recession (shortened as stability in TPR). This latter variable is a dichotomous one, as we have distinguished between having experienced any event of unemployment or unstable employment (fixed term contracts, paid Page 8 of 20 Verdetal. J Labour Market Res (2019) 53:4 individuals in each group.11 We present them briefly in the following lines. The first type, which we call the stable employment trajectory (44% of the cases in the sample) is characterised by job stability. Young people following this trajectory held stable employment for practically the entire time frame of our study. Therefore, the remaining employment situations have a negligible presence for those in this trajectory. Regarding their socio-demographic characteristics, young people with this trajectory are more likely to be older (30 to 34years of age), have secondary or higher educational levels, be women, have been born in Spain, and be the children of individuals in scientific, intellectual or liberal professions. The second type, which we call the precarious trajectory (26% of the cases) is characterised by unstable employment situations. Individuals with this trajectory experience long-term unemployment, which is combined with periods of inactivity and short odd jobs. In these cases, unemployment and temporality appear to be the common consequences of shared disadvantageous situations in the labour market, as suggested by Prieto and Pérez de Guzmán (2015). Young people with this type of trajectory are more likely to have low levels of education (at most havingonly finished compulsory education), to have been born outside of Spain, be under 30years of age and be the children of low-skilled workers. The third type, which we refer to as a temporary employment trajectory (20% of the cases) is mainly characterised by the prevalence of temporary employment. Individuals with this trajectory have suffered more unemployment than young people with a stable or nonsalaried trajectory, although in contrast to the precarious trajectory, their unemployment is short-term, and not long-term. The majority of individuals with this type of trajectory are under 30years of age and have parents in semi-skilled occupations and with middle levels of education (post-compulsory secondary studies). The last type, which we refer to as a non-salaried employment trajectory (accounting for only a bit more than 10% of our sample), has as its defining characteristic that individuals were either self-employed or business owners for the majority of the time examined. Almost the totality of individuals in this group were never unemployed during the entire period of the crisis studied. Regarding their socio-demographic characteristics, individuals with this trajectory are more likely to be older (from 30 to 34years of age), men, born in Spain, and children of managers and business owners. Table3 presents the coefficients from the two multinomial logistic regression models. The first includes only socio-demographic characteristics and the second adds the variables related to the trajectory prior to the recession (TPR). Comparing this second model with the first one, our initial finding is that, as expected, the explanatory power of the model increases when the three new variables related to the TPR of our sample of young people are incorporated. This can be seen in the higher R2 value of the second model. Nevertheless, even when introducing the TPR variables, most of the socio-demographic variables remain significant, particularly regarding the precarious and non-salaried trajectories. Only among the young people who experienced a trajectory of temporary employment does the introduction of TPR variables really introduce a relevant change in the significance of the socio-demographic variables. This suggests that sociodemographic characteristics might influence in a similar way the career path of young people with a precarious trajectory before and after the start of recession, and that the TPR has been more decisive in the likelihood of a young Table 2 Synoptic table summarizing themain characteristics oftheclusters Source: Elaborated by authors Stable employment trajectory Non-salaried employment trajectory Temporary employment trajectory Precarious trajectory Most characteristic events experienced during the recession (above-average) Stable employment Non-salaried employment Temporary employment Short odd jobs Unemployment Inactivity Experience of unemployment during the recession Never unemployed Never unemployed Short-term unemployment (less than a year) Long-term unemployment (one year or more) Socio-demographic characteristics of individuals in the cluster 30–34 years old Female Education: Post-compulsory secondary; higher Born in Spain Parents: Scientific, intellectual and liberal professions 30–34 years old Male Born in Spain Parents: Business owners and managers 20–29 years old Parents: Skilled and semiskilled workers 20–29 years old Education: Compulsory or less Born abroad Parents: Low-skilled workers 11 Tables5, 6 and 7 provide more detailed information on the distribution of these variables for every type of trajectory. Page 15 of 20 Verdetal. J Labour Market Res (2019) 53:4 Table 5 Number of individuals and average length of time in stable employment, non-salaried employment, unemployment, inactivity, shortodd jobs, temporary employment andinformal employment bytype oftrajectory Source: Calculated by authors based on the 2012 Catalan Youth Survey data Stable employment Non-salaried employment Unemploy-ment Inactivity Short odd jobs Temporary employment Informal employment Stable employment trajectory Mean 93.64 0.12 1.21 0.95 0.18 3.10 0.56 N 577 577 577 577 577 577 577 Sd 11.05 1.51 4.64 5.01 1.65 7.89 3.69 Precarious trajectory Mean 6.68 1.13 18.19 15.95 40.97 7.95 3.16 N 350 350 350 350 350 350 350 Sd 13.85 5.35 27.41 31.73 41.58 14.61 14.45 Non-salaried employment trajectory Mean 7.38 88.36 0.82 0.35 0.08 2.09 0.89 N 136 136 136 136 136 136 136 Sd 17.37 20.18 4.23 2.99 1.44 7.64 4.04 Temporary employment trajectory Mean 16.25 0.29 4.16 2.98 0.14 74.14 1.66 N 260 260 260 260 260 260 260 Sd 22.50 1.92 7.82 10.82 1.68 22.97 7.14 Total Mean 46.58 9.47 6.24 5.26 10.95 18.25 1.50 N 1322 1322 1322 1322 1322 1322 1322 Sd 44.26 27.62 16.55 18.50 27.97 31.03 8.59 Table 6 Frequencies oflength ofthelongest unemployment spell experienced bytype oftrajectory V Cramer = 0.359 (p < 0.001) Source: Calculated by authors based on the 2012 Catalan Youth Survey data Never unemployed Short-term unemployment (less thanayear) Long-term unemployment (one year ormore) Total Stable employment trajectory N 528 39 10 577 % column 51.3 25.5 7.2 43.6 Adjusted residual 10.5 − 4.8 − 9.2 Precarious trajectory N 194 40 116 350 % column 18.8 26.1 83.5 26.5 Adjusted residual − 11.8 − 0.1 16.1 Non-salaried employment trajectory N 127 6 3 136 % column 12.3 3.9 2.2 10.3 Adjusted residual 4.6 − 2.8 − 3.3 Temporary employment trajectory N 181 68 10 259 % column 17.6 44.4 7.2 19.6 Adjusted residual − 3.5 8.2 − 3.9 Total N 1030 153 139 1322 % column 100.0% 100.0% 100.0% 100.0% Page 16 of 20 Verdetal. J Labour Market Res (2019) 53:4 Table 7 Characteristics of age, sex, educational level, national origin, parents’ highest occupational status andcharacteristics ofthetrajectory prior totherecession bytype oftrajectory Stable employment trajectory Non-salaried employment trajectory Temporary employment trajectory Precarious trajectory Total Age group 20–24years old % row 21.9 5.8 26.3 46.0 100 Adjusted res. − 5.4 − 1.8 2.1 5.5 25–29years old % row 38.9 5.5 24.9 30.8 100 Adjusted res. − 2.5 − 4.2 3.5 2.6 30–34years old % row 50.6 14.2 15.1 20.1 100 Adjusted res. 5.7 5.1 − 4.6 − 5.8 Sex Female % row 47.4 7.3 19.4 26.0 100 Adjusted res. 2.6 − 3.4 − 0.3 − 0.4 Male % row 40.3 12.9 19.9 26.9 100 Adjusted res. − 2.6 3.4 0.3 0.4 Education Compulsory or less % row 30.4 10.9 21.2 37.5 100 Adjusted res. − 6.9 0.5 0.9 6.6 Post-compulsory secondary % row 50.7 7.8 18.2 23.3 100 Adjusted res. 2.8 − 1.7 − 0.8 − 1.3 Higher % row 50.3 11.2 19.4 19.1 100 Adjusted res. 4.3 0.9 − 0.3 − 5.2 National origin Born in Spain % row 47.1 11.0 20.2 21.8 100 Adjusted res. 6.1 2.1 1.1 − 9.3 Immigrant % row 23.9 6.1 16.8 53.3 100 adjusted res. − 6.1 − 2.1 − 1.1 9.3 Parents’ highest occupational status Business owners and managers % row 36.4 24.8 14.0 24.8 100 Adjusted res. − 1.6 5.5 − 1.7 − 0.5 Scientific, intellectual and liberal professions % row 52.9 8.7 14.5 23.9 100 Adjusted res. 2.5 − 0.7 − 1.7 − 0.8 Middle management and technical employees % row 43.1 13.9 18.1 25.0 100 Adjusted res. 0 1.5 − 0.6 − 0.5 Skilled and semi-skilled workers % row 43.4 8.5 23.1 25.1 100 Adjusted res. 0.3 − 2.6 3.2 − 1.5 Page 17 of 20 Verdetal. J Labour Market Res (2019) 53:4 Table 7 (continued) Stable employment trajectory Non-salaried employment trajectory Temporary employment trajectory Precarious trajectory Total Low-skilled workers % row 37.1 5.7 16.4 40.7 100 Adjusted res. − 1.5 − 1.9 − 1.1 4.0 Stability in TPR Experience of unemployment or unstable employment % row 40.4 7.5 22.4 29.7 100 Adjusted res. − 5.4 − 7.5 5.6 6.1 No experience of unemployment or unstable employment % row 60.6 24.4 5.6 9.4 100 Adjusted res. 5.4 7.5 − 5.6 − 6.1 Highest occupational status acquired TPR Low-skilled job % row 17.6 11.0 22.0 49.5 100 Adjusted res. − 5.5 0.0 0.4 6.1 Skilled and semi-skilled job % row 44.3 8.6 19.1 28.0 100 Adjusted res. − 0.8 − 2.8 − 1.0 3.9 High-skilled job % row 51.8 14.0 21.3 12.8 100 Adjusted res. 3.7 2.8 0.8 − 7.2 Length TPR 0–2years % row 33.9 6.1 27.7 32.3 100 Adjusted res. − 4.5 − 3.1 4.7 3.0 3–5years % row 48.0 11.3 15.5 25.1 100 Adjusted res. 1.9 0.7 − 2.3 − 0.6 6years or more % row 47.3 12.3 16.9 23.5 100 Adjusted res. 2.4 2.2 − 2.2 − 2.2 Total % row 43.6 10.3 19.7 26.4 100 Percentages and adjusted residuals Calculated by authors based on the 2012 Catalan Youth Survey data Page 18 of 20 Verdetal. J Labour Market Res (2019) 53:4 Table 8 Multinomial logistic regression forthesubpopulation of20–29years old (N M1 = 584 M2 = 545) Reference: stable employment trajectory M1 M2 B OR B OR Precarious trajectory Constant 0.597 (0.47) − 1.898 (0.69) ** Female 0.357 (0.23) 1.429 0.368 (0.26) 1.445 Male 0 1 0 1 20–24 years old 0.821 (0.26) 2.273** 0.759 (0.33) 2.136* 25–29 years old 0 1 0 1 Born in Spain − 1.295 (0.30) 0.274* − 0.561 (0.35) 0.571 Born abroad 0 1 0 1 Compulsory or less 0.314 (0.28) 1.369 0.397 (0.35) 1.487 Post-compulsory secondary 0.332 (0.29) 1.394 0.336 (0.34) 1.399 Higher education 0 1 0 1 Parents: business owners and managers 0.482 (0.48) 1.619 0.659 (0.53) 1.933 Parents: scientific. intellectual and liberal professions − 0.238 (0.48) 0.788 − 0.223 (0.52) 0.800 Parents: middle management and technical employees 0.337 (0.50) 1.400 0.933 (0.56) 2.541 Parents: skilled and semi-skilled workers − 0.263 (0.35) 0.769 − 0.347 (0.38) 0.707 Parents: low-skilled workers 0 1 0 1 Highest occup. status TPR: low-skilled 2.060 (0.55) 7.845* Highest occup. status TPR: semi-skilled 0.645 (0.30) 1.906 Highest occup. status TPR: high-skilled 0 1 Experienced unemployment or unstable employment in TPR 1.692 (0.41) 5.429* Not unemployment or unstable employment in TPR 0 1 Length TPR − 0.109 (0.06) 0.897 Non-salaried employment trajectory Constant − 4.683 (1.66) ** − 8.227 (2.01) ** Female − 1.413 (0.61) 0.243* − 1.216 (0.67) 0.296 Male 0 1 0 1 20–24 years old 0.238 (0.48) 1.269 2.116 (0.75) 8.294* 25–29 years old 0 1 0 1 Born in Spain 2.052 (1.46) 7.783 2.023 (1.49) 7.562 Born abroad 0 1 0 1 Compulsory or less 1.539 (0.70) 4.659* − 0.031 (0.84) 0.970 Post-compulsory secondary 1.733 (0.72) 5.657* 1.142 (0.77) 3.134 Higher education 0 1 0 1 Parents: business owners and managers 1.224 (0.77) 3.400 2.109 (0.89) 8.241 Parents: scientific, intellectual and liberal professions − 0.198 (0.96) 0.821 0.844 (1.05) 2.325 Parents: middle management and technical employees − 0.086 (0.96) 0.918 1.402 (1.08) 4.065 Parents: skilled and semi-skilled workers − 0.356 (0.65) 0.701 0.222 (0.72) 1.248 Parents: low-skilled workers 0 1 0 1 Highest occup. status TPR: low-skilled 2.878 (0.87) 17.776* Highest occup. status TPR: semi-skilled 1.087 (0.66) 2.964 Highest occup. status TPR: high-skilled 0 1 Experienced unemployment or unstable employment in TPR 0.463 (0.52) 1.589 Not unemployment or unstable employment in TPR 0 1 Length TPR 0.441 (0.15) 1.554* Temporary employment trajectory Constant − 0.454 (0.52) − 3.153 (0.81) ** Female 0.275 (0.24) 1.316 0.218 (0.26) 1.244 Male 0 1 0 1 Page 19 of 20 Verdetal. J Labour Market Res (2019) 53:4 Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Received: 28 February 2017 Accepted: 14 February 2019 References Alós, R.: Segmentación de los mercados de trabajo y relaciones laborales. El sindicalismo ante la acción colectiva. Cuadernos de Relaciones Laborales 26(1), 123–148 (2008) Alós, R.: Desempleo y empleo durante la crisis. In: Miguélez, F. (coord.) Diagnóstico socioeconómico sobre las políticas de empleo en España, 2012–2014. Universitat Autònoma de Barcelona, Bellaterra (Cerdanyola del Vallès). http://ddd.uab.cat/recor d/14288 4 (2015) Alós, R., Lope, A.: El desempleo y sus consecuencias: vulnerabilidad y riesgo de exclusión social. In: Torres, C. (ed.) España 2015. Situación Social. Centro de Investigaciones Sociológicas, Madrid (2015) Baert, S., Bart, C., Verhaest, D.: Overeducation at the start of the career: stepping stone or trap? Labour Econ. 25, 123–140 (2013) Bande, R., Karanassou, M.: Spanish regional unemployment revisited: the role of capital accumulation. Reg. Stud. 48(11), 1863–1883 (2014) Banyuls, J., Miguélez, F., Recio, A., Cano, E., Lorente, R.: The transformation of the employment system in Spain: towards a mediterranean neoliberalism. In: Bosch, G., Lehndorff, S., Rubery, J. (eds.) European employment models in flux. A comparison of institutional change in Nine European Countries, pp. 247–269. Palgrave-Macmillan, New York (2009) Barbieri, P., Scherer, S.: Labour market flexibilization and its consequences in Italy. Eur. Sociol. Rev. 25(6), 677–692 (2009) Bell, D.N., Blanchflower, D.G.: Youth unemployment in Europe and the United States. Nordic. Econ. Policy Rev. 1, 11–38 (2011) Belvis, F.X., Benach, J.: Educació i estabilitat laboral a Catalunya. Mobilitat entre ocupació estable, inestable i no ocupació, 2002–2012. Fundació Jaume Bofill, Barcelona. http://www.fbofi ll.cat/publi cacio ns/educa cioi-estab ilita t-labor al-catal unya-0 (2014) Bentolila, S., Dolado, J.J., Jimeno, J.F.: Two-tier employment protection reforms: the Spanish experience, CESifo DICE Report 4/2008. http:// www.cesif o.de (2008) Bentolila, S., Cahuc, P., Dolado, J.J., Le Barbanchon, T.: Two-tier labour markets in the Great Recession: France versus Spain. Econ. J. 122, F155–F187 (2012) Bernardi, F., Garrido, L.: Is there a new service proletariat? Post-industrial employment growth and social inequalities in Spain. Eur. Sociol. Rev. 24(3), 299–313 (2008) Berton, F., Devicienti, F., Pacelli, L.: Are temporary jobs a port of entry into permanent employment? Evidence from matched employer– employee. Int. J. Manpower 32(8), 879–899 (2011) Bruno, G.S.F., Caroleo, F.E., Dessy, O.: Stepping stones versus dead end jobs: exits from temporary contracts in Italy after the 2003 Reform (IZA discussion Paper No. 6746). http://hdl.handl e.net/ (2012) Cachón, L.: En la ‘España inmigrante’: entre la fragilidad de los inmigrantes y las políticas de integración. Papeles CEIC 1, 1–35 (2009) Caparrós, A., Navarro, M.L.: Temporalidad, segmentación laboral y actividad productiva: ¿existen diferencias regionales? Estadística Española 50(168), 205–245 (2008) Casal, J., García, M., Merino, R., Quesada, M.: Changes in forms of transition in contexts of informational capitalism. Papers. Revista de Sociologia 79, 195–223 (2006) Castelló, L., Bolíbar, M., Barranco, O., Verd, J.M.: Treball. Condicions en el mercat de treballi trajectòries laborals de la joventut catalana. In: Serracant, Table 8 (continued) Reference: stable employment trajectory M1 M2 B OR B OR 20–24 years old 0.569 (0.28) 1.766 0.125 (0.35) 1.134 25–29 years old 0 1 0 1 Born in Spain − 0.364 (0.35) 0.695 0.354 (0.40) 1.425 Born abroad 0 1 0 1 Compulsory or less 0.394 (0.28) 1.482 0.619 (0.35) 1.857 Post-compulsory secondary 0.144 (0.31) 1.155 0.176 (0.35) 1.193 Higher education 0 1 0 1 Parents: business owners and managers 0.026 (0.56) 1.026 − 0.176 (0.60) 0.838 Parents: scientific, intellectual and liberal professions − 0.255 (0.54) 0.775 − 0.651 (0.58) 0.522 Parents: middle management and technical employees 0.309 (0.54) 1.363 0.819 (0.60) 2.268 Parents: skilled and semi-skilled workers 0.037 (0.38) 1.038 − 0.021 (0.41) 0.979 Parents: low-skilled workers 0 1 0 1 Highest occup. status TPR: low-skilled 1.781 (0.56) 5.939* Highest occup. status TPR: semi-skilled 0.331 (0.30) 1.393 Highest occup. status TPR: high-skilled 0 1 Experienced unemployment or unstable employment in TPR 2.680 (0.55) 14.591* Not unemployment or unstable employment in TPR 0 1 Length TPR − 0.186 (0.06) 0.830* Influence of socio-demographic characteristics of respondents and characteristics of the trajectory experienced prior to the recession on the type of labour trajectory. Beta coefficients, standard error (in brackets) and Odds Ratio R2 Nagelkerke M1 = 0.163; M2 = 0.322 Source: Calculated by authors based on the 2012 Catalan Youth Survey data **p < 0.01, *p < 0.05 Page 20 of 20 Verdetal. J Labour Market Res (2019) 53:4 P. (ed.) Enquesta a la joventut de Catalunya 2012, vol. 1, pp. 117–218. Barcelona, Generalitat de Catalunya (2013) Choudhry, M.T., Marelli, E., Signorelli, M.: Youth unemployment and the impact of financial crisis. Int. J. Manpower 33(1), 76–95 (2012) De la Rica, S., Anghel, B.: Los parados de larga duración en España en la crisis actual. Documento de trabajo 185/2014. Fundación Alternativas, Madrid (2014) Debels, A.: Transitions out of temporary jobs: consequences for employment and poverty across Europe. In: Muffels, R.J.A. (ed.) Flexibility and employment security in Europe, pp. 51–77. Labour markets in Transition, Edward Elgar, Cheltenham (2008) Dietrich, H., Möller, J.: Youth unemployment in Europe: business cycle and institutional effects. Int. Econ. Econ. Policy 13(1), 5–25 (2016) Doeringer, P., Piore, M.: Internal labor markets and manpower analysis. Heath and Co, Lexington (1971) Edwards, R.C.: The social relations of production in the firm and labour market structure. Polit. Soc. 5, 83–108 (1975) Eichhorst, W., Marx, P., Wehner, C.: Labor market reforms in Europe: towards more flexicure labor markets? J. Labour Market Res. 51, 3 (2017) Erikson, R.: Social class of men, women and families. Sociology 18(4), 500–514 (1984) Freeman, R.B., Wise, D.A. (eds.): The youth labor market problem: its nature, causes and consequences. University of Chicago Press/NBER, Chicago (1982) García-Pérez, J.I., Muñoz-Bullón, F.: Transitions into permanent employment in Spain: an empirical analysis for young workers. Br. J. Ind. Relat. 49(1), 103–143 (2011) Hair Jr., J.F., Black, W.C., Babin, B.J., Anderson, R., Tatham, R.E.: Multivariate data analysis, 7th edn. Prentice Hall, New Jersey (2009) Häusermann, S., Schwander, H.: Varieties of dualization? Labor market segmentation and insider-outsider divides across regimes. In: Emmenegger, P., Häusermann, S., Palier, B., Seeleib-Kaiser, M. (eds.) The age of dualization. The changing face of inequality in deindustrializing societies, pp. 27–51. Oxford University Press, Oxford (2012) Hernanz, V.: El trabajo temporal y la segmentación: un estudio de las transiciones laborales. Consejo Económico y Social, Madrid (2003) ILO: Spain growth with jobs. International Labour Office, Geneva (2014) ILO: World employment and social outlook 2016: trends for youth. International Labour Office, Geneva (2016) Kaliski, S.F.: Why must unmeployment remain so high? Can. Public Pol. 10(2), 127–141 (1984) Korupp, S.E., Ganzeboom, H.B.G., Van Der Lippe, T.: Do mothers matter? A comparison of models of the influence of mothers and fathers educational and occupational status on children’s educational attainment. Qual. Quant. 36, 17–42 (2002) Lefresne, F.: Les jeunes et l’emploi. La Découverte, Paris (2003) López-Andreu, M., Verd, J.M.: Employment instability and economic crisis in Spain: what are the elements that make a difference in the trajectories of younger adults? Eur. Soc. 18(4), 315–335 (2016) López-Bazo, E., Montellón, E.: The regional distribution of unemployment: what do micro-data tell us? Pap. Reg. Sci. 92(2), 383–405 (2013) Meroni, C.E., Vera-Toscano, E.: The persistence of overeducation among recent graduates. Labour Econ. 48, 120–143 (2017) Mertens, A., Gash, V., McGinnity, F.: The cost of flexibility at the margin. Comparing the wage penalty for fixed-term contracts in Germany and Spain using Quantile Regression. Labour 21(4/5), 637–666 (2007) Muñoz-de-Bustillo, R., Esteve, F.: The neverending story. Labour market deregulation and the performance of the Spanish labour market. In: Piasna, A., Myant, M. (eds.) Myths of employment deregulation: how it neither creates jobs nor reduces labour market segmentation, pp. 61–80. ETUI, Brussels (2017) O’Higgins, N.: This Time It’s Different? Youth labour markets during ‘The Great Recession’ (IZA discussion paper, No. 6434). Institute for the Study of Labor, Bonn (2012) O’Reilly, J., Eichhorst, W., Gábos, A., Hadjivassiliou, K., Lain, D., Leschke, J., McGuinness, S., Kureková, L.M., Nazio, T., Ortlieb, R., Russell, H., Villa, P.: Five characteristics of youth unemployment in Europe: flexibility, education, migration, family legacies, and EU policy. SAGE Open 5(1), 1–19 (2015) Osterman, P.: Getting started: the youth labor market. MIT Press, Camdridge (1980) Pastore, F.: The youth experience gap. Explaining national differences in the school-to-work transition. Springer, Heidelberg (2015) Pastore, F.: Why so slow? The school-to-work transition in Italy (IZA discussion paper, No. 10767). Institute for the Study of Labor, Bonn (2017) Pastore, F., Giuliani, L.: The determinants of youth unemployment. A panel data analysis (CRISEI, discussion paper, No.2). University of Naples Parthenope, Naples (2015) Prieto, C., Arnal, M., Caprile, M., Potrony, J.: La calidad del empleo en España: una aproximación teórica y empírica. Ministerio de Trabajo y Seguridad Social, Madrid (2009) Prieto, C., Pérez de Guzmán, S.: La precarización del empleo en el marco de la norma flexible-empresarial de empleo. In: Torres, C. (ed.) España 2015. Situación social. Centro de Investigaciones Sociológicas, Madrid (2015) Quintini, G., Martin, J., Martin, S.: The changing nature of the school-to-work transition process in OECD countries (IZA discussion paper, No. 2582). Institute for the Study of Labor, Bonn (2007) Rocha, F.: La crisis económica y sus efectos sobre el empleo en España”. Gaceta Sindical 19, 67–89 (2012a) Rocha, F.: Youth unemployment in Spain. Situation and policy recommendations. Friedrich-Ebert-Stiftung, Berlin (2012b) Ryan, P.: The school-to-work transition. A cross-national perspective. J. Econ. Lit. 39(1), 1–34 (2001) Sala, H., Trivín, P.: Labour market dynamics in Spanish regions: evaluations asymmetries in troublesome times. SERIEs 5, 197–221 (2014) Salido, O.: La movilidad ocupacional de las mujeres en España. Por una sociología de la movilidad femenina. Centro de Investitgaciones Sociológicas, Madrid (2001) Scherer, S.: Stepping-stones or traps? The consequences of labour market entry positions on future careers in West Germany, Great Britain and Italy. Work Employ Soc. 18(2), 369–394 (2004) Serracant, P.: Changing youth? Continuities and ruptures in transitions into adulthood among Catalan young people. J. Youth Stud. 15(2), 161–176 (2012) Tomlinson, M., Walker, R.: Labor market disadvantage and the experience of recurrent poverty. In: Emmenegger, P., Häusermann, S., Palier, B., SeeleibKaiser, M. (eds.) The age of dualization. The changing face of inequality in deindustrializing societies, pp. 52–70. Oxford University Press, Oxford (2012) Verd, J.M., López-Andreu, M.: La inestabilidad del empleo en las trayectorias laborales Un análisis cuantitativo. Rev. Esp. Investig. Soc. 138, 135–148 (2012) Verick, S.: The impact of the global financial crisis on labour markets in OECD countries: why youth and other vulnerable groups have been hit hard. In: Islam, I., Verick, S. (eds.) From the great recession to labour market recovery, pp. 119–145. Palgrave Macmillan, London (2011) Ward, J.H.: Hierarchical grouping to optimize an objective function. J. Am. Stat. Assoc. 58, 236–244 (1963)