Leaving home across the recent cohorts in Italy: does economic vulnerability due to labour market status matter?

We focus on the process of leaving home among young Italians and aim to discern: (a) the extent to which the potential economic vulnerability associated with certain individual labour market conditions affects the pathway of leaving the parental home; (b) whether the impact of such conditions has changed over generations of young adults. Data come from the pooling of two cross-sectional rounds of the ‘Families and Social Subjects’ survey conducted by the Italian National Institute of Statistics in 2009 and 2016. Specifically, we consider young adults who at the time of interview were between the ages of 20 and 44, for a total of 22,218 respondents. We employ discrete-time event history models, including a competing risks approach. Our findings suggest that labour market status differs in its influence according to reason for exit and gender. While we observe a clearly negative impact of being out of the labour market for both union and non-union related reasons for men, this negative impact is observed only for non-union-related motives for women. Meanwhile, economic uncertainty connected with precarious jobs does not necessarily have a negative impact on the risk of leaving parental home, both for union and non-union related reasons. Furthermore, we see signs of differences across cohorts relative to this latter set of motives for men. The results highlight the importance of considering reasons for exit in analyses of departure from the parental home.


Introduction
Examining departure from the parental home is of crucial importance, as this is one of the salient markers in the transition to adulthood (TTA) process.Indeed, the way this step is experienced can have important consequences for young adults' life course outcomes.For example, late exit from the parental home may contribute to the postponement of other transitions (i.e.forming a stable cohabiting partnership or having children), hence generally prolonging the transition to adult roles (Furstennerg 2010;Liefbroer & Toulemon, 2010;Krahn et al., 2018).An early transition out of the parental home may instead be associated with earlier experiences of other TTA events (i.e.exit from education, entry into work or forming a family) that could prevent young people from acquiring adequate human capital for later in life (Osgood et al., 2005;Schwanitz, 2017).
Given the significance of this life event, a rich literature explores the determinants of the decision to leave the parental home in Western European societies (e.g.Aassve et al., 2002Aassve et al., , 2013aAassve et al., , 2013b;;Billari & Liefbroer, 2010;Chiuri & Del Boca, 2010;Corijn & Klijzing, 2001;Iacovou, 2010;Mazzuco & Ongaro, 2009;Mulder & Clark, 2000;Schwanitz et al., 2017;Tosi, 2017).Studies show that among the numerous (individual, family, contextual) factors influencing the decision to establish an independent household, young adults' economic self-sufficiency plays an important role.In particular, empirical research demonstrates that young people who can rely on an own income or stable/fulltime employment are more likely to achieve residential emancipation from the parental home (Bertolini et al., 2017;Iacovou, 2010;Mulder & Clark, 2000;Whittington & Peters, 1996).However, scholars also document that the impact of personal economic resources can vary depending on the institutional and cultural context as well as gender and reason for leaving home (Aassve et al., 2002;Iacovu 2010;Schwanitz et al., 2017).
This paper assesses young Italians' departure from the parental home from a cohort perspective.Specifically, we analyse: (a) the role played by individual labour market conditions in the process of leaving the parental home with the assumption that some of these conditions (out of labour market or insecure job) may be a proxy of personal economic vulnerability.We further differentiate by reason of exit (union or non-unionrelated motives); (b) the extent to which the impact of economic vulnerability has changed among the more recent cohorts (i.e.those born during the 1980s and early 90s, the so-called millennials 1 ) compared to previous cohorts (those born between 1965 and 1979).
Italy offers a particularly interesting case for study.Young Italians traditionally tend to leave home at later ages compared to their counterparts in Central or Northern European countries (Aassve et al., 2002;Aassve et al., 2013aAassve et al., , 2013b;;Corijn and Kleijzing 2001;Iacovu 2010) and mainly do so for union-related reasons (usually marriage).A historical interaction between cultural factors and a particular institutional framework (Ongaro, 2005) underlies this pattern and the result is a familistic welfare regime (Dalla Zuanna, 2001;Dalla Zuanna & Micheli, 2004;Esping-Andersen, 1999), whereby the coresidence of young adults with parents until they are ready to form a new family of their own is viewed as socially acceptable.
However, recent cohorts of Italians, especially those born after 1980, have been exposed to cultural and (structural and conjunctural) economic changes, which may have affected their pathways to residential autonomy.Like previous generations, millennials live in a context characterized by persistent familistic welfare.They too have similarly been exposed to the effects of worsening work conditions that can be traced back to the structural labour market reform of the 1980s and 1990s (Vignoli et al., 2016).Yet, differently from previous cohorts, they experienced conjunctural economic crises due to both the 2008 Great Recession and the 2010 current sovereign debt crisis.
Since the 1990s, young Italians transitioning into adulthood have necessarily taken on jobs that are highly vulnerable to unemployment (temporary contracts) or unstable work at higher risk of economic uncertainty (self-employment).Economic crises of the early 2000s further aggravated the vulnerability of young adults, strongly increasing youth unemployment, low entry salaries, and temporary/unstable jobs.Millennials are furthermore amongst the first to have more extensively experienced: (a) the cultural and behavioural changes associated with the spread of the Second Demographic Transition (non-marital unions, out-of-wedlock childbearing and divorce); (b) the diffusion of a 'globalized' youth culture that emphasizes individual rights, post-materialistic values and gender equity (Cepa & Furstenberg, 2021;Leccardi & Ruspini, 2006), fostered by the spread of social networks and internet connections (Digital 2022) and by increasing international mobility.All these changes may have affected the association between individual economic vulnerability and leaving home and hence millennials' paths to residential autonomy.For example, the spread of a less long-term commitment transition (i.e.cohabitation) could have reduced the negative impact of individual economic uncertainty on exit for union-related reasons.Similarly, the greater confidence of millennials in dealing with a precarious economic context may have made them more willing to exit for non-union-related reasons in the case of an uncertain job.
Though a number of studies explore the role of economic self-sufficiency (measured by either income or employment status) in youths' propensity to leave the parental home, relatively few consider both type of economic vulnerability and type of exit from the parental home (Bertolini et al., 2017), and none have adopted a cohort perspective.Moreover, for Italy, relevant empirical analyses stop at the cohorts born in the early 1970s (Billari et al., 2002;Ongaro, 2001).Our investigation thus sheds greater light not only on the interrelationship between different kinds of economic vulnerability associated with labour market conditions and reasons for exiting the parental home, but alsoin assessing cohorts born between 1965 and 1996-on the ways that young people today depart the parental home in a 'traditional' society that is experiencing important societal and cultural changes.To this end, we use retrospective data from the pooling of two rounds of the 'Families and Social Subjects' survey conducted by the Italian National Institute of Statistics (ISTAT) in 2009 and 2016, and apply event history models (either as a single 'destination' or a competing risk version), distinguishing between men and women.
The rest of the paper proceeds as follows.Section "Youth economic self-sufficiency and leaving home" reviews the literature on the relationship between economic conditions and leaving home.Section "The Italian setting and hypothese" describes the specificities of the Italian setting and sets forth our hypotheses.Section "Data, selection of the sample, preliminary descriptions, and statistical implementation" presents the data and statistical models followed, in section "Results", by our findings.Section "Conclusion and discussion" concludes with a discussion of the results.

Youth economic self-sufficiency and leaving home
A rich empirical literature explores the importance of individual economic resources in young adult household formation, with various studies looking specifically at the role of personal income and occupational status in shaping youths' departure from home.
Early research based on data from high-income countries mostly documents a positive relationship between economic self-sufficiency and leaving home (for a more complete review of this literature, see Aassve et al., 2002).For instance, Ermisch (1999) reports that young Brits with high levels of income have a higher probability of leaving the parental home and a lower probability of returning.Several studies on young adults in the US (Avery et al., 1992;Mulder & Clark, 2000) similarly show that own income (and especially a high level of personal income) has a significant positive impact on exit from the parental home.Going back to European societies, the studies focusing on occupational status document that being employed plays an important role in accelerating the transition to independent living.Wagner and Huinink (1991), for instance, observe this result for West Germany across birth cohorts.Nilsson and Strandh (1999) meanwhile report that being continuously employed has a strong positive impact on the propensity to leave home for young adults in Sweden, whereas experiencing labour market adversities tends to increase the propensity to return home.Finally, some works on Mediterranean countries (Aassve et al. 2001;Billari et al., 2002) finds that-more than income itself-being employed is an important prerequisite for leaving the parental home for both Italian and Spanish young men; in contrast, employment status has no impact on the likelihood of becoming independent of their parents for women.
The literature also demonstrates the importance of distinguishing between different pathways out of the parental home.Departure is, in fact, closely connected with union formation, which complicates analyses.Whittington and Peters (1996) report that personal financial resources are more important for young American men than women, especially relative to partnership decisions.Hughes's (2003) findings support these results.In her examination of the relationship between labour and housing market conditions and the different living arrangements of young adults in the US, she observes that higher incomes are particularly associated with marriage compared to other situations (especially residing with parents or roommates), and this is true for both men and women.Iacovu's (2010) multinational European comparative study shows that young person's own income has a strong positive association with leaving home for unionrelated reasons (for both men and women), while a weaker association appears for exits for personal autonomy, and even more so for education motives.Looking at Spain, Del Rey et al. (2022) find that for men, financial situation is an important factor when leaving home, regardless of reason for departure, whereas for women labour market trajectories are positively associated with leaving home for cohabitation or for non-union related reasons, but negatively associated with leaving home to marry.
These empirical results support the theory that the impact of economic conditions may vary by reasons for exit.According to economic theory, individual financial vulnerability is likely to discourage youth from making long-term binding commitments such as partnerships and, particularly, marriage (Becker, 1981;Blossfeld et al., 2005;Schoon and Binner, 2017).The same studies also highlight the importance of considering gender differences in leaving home, especially exiting for union-related reasons.Gender role theories suggest, indeed, that women may be less sensitive than men to economic and job insecurity, not only because they may prefer to invest primarily in "marriage careers" (Hakim, 2001) but also because they may be more likely than men to react to negative or unsatisfactory employment prospects by choosing alternative (family) careers (Friedman et al., 1994).Moreover, in societies where the traditional male-breadwinner model predominates, the impact of economic conditions on union transitions is more gender specific (Blossfeld et al., 2005;Del Rey et al., 2022;Leschke & Jepsen, 2012;Oppenheimer, 1988).Namely, as primary income earners, men are expected to have a job and be financially autonomous before establishing their own family, while for women, having a job or being financially independent does not seem to be a prerequisite.
Studies furthermore suggest that the effect of personal economic resources may vary depending on a country's welfare regime.Generous welfare policies toward young adults can support the transition to financial autonomy, thus reducing the impact of income or job stability on the likelihood of leaving the parental home.Using the Esping-Andersen typology (1999), several papers demonstrate that in the 'familistic' Southern European welfare state, personal economic self-sufficiency (measured as either income or employment status) is an important factor in young adults' decision to leave home, while it is negligible in the 'social-democratic' regimes of Nordic countries (Aassve et al., 2002(Aassve et al., , 2006;;Billari, 2004).These results echo the findings of Arundell and Lennartz (2017, p. 13) who-in analysing the relationship between welfare regimes and several individual life-course turning point shocks-'provide empirical evidence of a supplementary "welfare regime effect", whereby there is a higher propensity to return to parental co-residence in the more familistic Southern European context' when shifts to unemployment or inactive conditions occur.
To summarize, although some effects may differ by reasons for exit, gender or welfare regime, the empirical literature suggests that individual economic vulnerability-particularly a lack of personal income due to labour market exclusion (i.e.being unemployed or inactive)-may deter young people from pursuing residential independency.
Somewhat less clear is whether housing independence from parents may also vary depending on employment insecurity in the form of a (paid) non-permanent job.This is, however, a fundamental question, as young adults have increasingly faced a market that comprised non-standard jobs (temporary, atypical or non-formal types of contracts), largely due to labour deregularization starting in the 1980s (Mills et al., 2005).It may be that-irrespective of the amount of pay-foreseen variation in income connected with jobs with high degrees of economic uncertainty deters young adults from establishing their own household (Becker et al., 2010;Fernandes et al., 2008;Mills and Blossfeld, 2003;Mills et al., 2005;Oppenheimer et al., 1997).Yet, while the empirical literature finds a negative association with job insecurity for certain TTA events (e.g. for first parenthood, see, among others, Barbieri et al., 2014;Dupray & Pailhé, 2017;Vignoli et al., 2020), relatively few studies detect a negative effect of non-standard forms of employment on the probability of residential independence.While an analysis of British youth documents that temporary or part-time (permanent) jobs do have a negative effect on the likelihood of housing autonomy for women (Gousia et al., 2021;Gebel, 2017) shows that in Germany, having a temporary contract or being self-employed with respect to having a permanent job, does not reduce, for either men or women, the likelihood of exiting the parental home.In Italy as well, objective job insecurity stemming from a temporary contract does not seem to have a statistically significant impact on the risk of housing autonomy among young adults, and this finding refers to both men and women and to exits for both union and non-union-related reasons (Bertolini et al., 2017).Contrary then to what one might expect (Blossfeld et al., 2005(Blossfeld et al., , 2011)), temporary employment may not have a significant impact on leaving home, even in countries where the welfare system is less generous in supporting youth in their transition to residential autonomy (Ranci et al., 2014).Broadly, the empirical evidence suggests that job precariousness is not detrimental to housing autonomy.

The Italian setting and hypotheses
In Italy (and in other Southern European countries) leaving home traditionally occurs at later ages than in Northern or Western European countries and mainly for unionrelated reasons (Aassve et al., 2013a(Aassve et al., , 2013b;;Billari et al., 2002;Corijn & Klijzing, 2001).For example, in the early 1990s, 32% of Italian men and 20% of women had never left home at the age of 30, compared to 2% and 1% in Sweden, 9% and 5% in France and 11% and 5% in the UK (Billari et al. 2002).Ten years later, relatively little had changed (Assve et al., 2013b): in Italy 57% of young adults were living with their parents at age 25-29 compared to 19% in the UK, 12% in France, and 3% in Sweden.Recent data from Eurostat confirm the persistence of this trend. 2 recursive interaction between long-term cultural continuities and institutional and economic factors underlies this pattern (Ferrari et al. 2014;Mazzuco et al., 2006;Ongaro, 2005).Italy is a country characterized by 'strong family ties' (Reher, 1998), entailing a reliance on intergenerational support rather than welfare transfers (Dalla Zuanna & Micheli, 2004).In other words, it is largely the family of origin that provides emotional and material aid to young people leaving home (Bertolini, 2011;Mencarini & Tanturri, 2006).Strong intergenerational relationships are reinforced by an institutional framework that provides weak support for early departure from the parental home (see, among others, Barbieri, 2011;Rosina et al., 2007).Consider, for example, higher education.In Italy, young adults are encouraged to attend local public universities, which often means that rather than residing in on-campus accommodation, they continue to live with their parents during their studies.The housing market further complicates things.Affordable rented accommodation is scarce due to a rigid homeownership regime and lack of a well-functioning mortgage system, making it difficult for young people to establish an own household (see, for example, Mulder & Billari, 2010or Modena & Rondinelli, 2016).
Conservative family values may also help to explain why the traditional path to housing autonomy is through union formation (especially marriage; for a description of differences among European countries to this regard, see Corijn & Klijzing, 2001).The family patterns associated with the Second Demographic Transition (SDT) began relatively late in Italy, with the diffusion of cohabitation, in particular, starting only in the late 1990s (Pirani & Vignoli, 2016;Vignoli et al., 2016).Even today, consensual unions are comparatively less common than in other European countries (Aisp, 2015;Billari et al., 2002;Nazio & Blossfeld, 2003) and this could partially explain why Italian youth postpone the exit for union.Furthermore, leaving home early or departing for non-union-related reasons are also much less socially acceptable than in other central and northern European countries (Tosi, 2017).
The structure of the labour market is another important factor shaping late exit from the parental home.The process of labour market flexibilization, which started in Italy during the 1980s and saw various laws and reforms during the 1990s and early 2000s, culminated in the spread of new flexible and temporary contracts (Barbieri & Scherer, 2009) that increased, in particular, the economic vulnerability of young people and women (Barbieri, 2011;ISTAT, 2014a;Vignoli et al., 2016).Objective job insecurity may thus represent an additional reason for the postponement of departure from the parental home among young adults born after the 1960s, especially as marriage has traditionally been the path to housing autonomy.
The above aspects intersect with relatively traditional gender roles in Italy, which manifest in persistent inequalities in both the labour market and family life (Dotti Sani, 2018;ISTAT, 2019).
In light of the above-described features of the Italian context and findings of previous literature, we formulate the following first set of hypotheses on the role of employment status in shaping the leaving home pathway: HP1: Labour market exclusion has a negative effect on transition out of the parental home for union-related reasons for men and, to a lesser extent, for women.Basically, Italy is a relatively conservative country in terms of gender roles, where the male breadwinner model persists and men are often expected to be the family's main provider of economic resources.Economic aspects may be less important for women, since they tend to be more involved than men in family tasks.HP2: Labour market exclusion has a negative effect on the risk of leaving home for non-union-related reasons.Since having a minimum of personal income is a prerequisite for all who intend to establish an autonomous household in a country where the welfare system does not support individuals without job (even in situations of job loss) or offset housing costs, the effect refers to both men and women.HP3: The persistence of traditional gender roles in Italy means that precarious employment (temporary contracts and self-employment) has a negative effect on transition out of the parental home for union-related reasons for men and, to a lesser extent, for women.This expectation aligns with previous theoretical frameworks (see, for example, Mills et al., 2005), though the only Italian study on this topic shows no effect (Bertolini et al., 2017).Since the potential economic vulnerability associated with self-employment may be less strong than that associated with temporary jobs, we might expect that its negative effects in such cases are weaker.HP4: Given the inconsistent results of the theoretical and empirical literature, it is not easy to formulate a clear hypothesis on the impact of precarious employment (temporary contracts and self-employment) on exit for non-union-related reasons.Following previous empirical evidence, we might assume that it has no effect on this type of departure.However, given the particularities of the Italian setting (labour market and welfare protection), we expect that precarious employment has a weak negative effect for both men and women.Again, the negative effects for self-employment may be less strong than for temporary jobs.
The role of economic vulnerability hypothesized above could, however, be different for the more recent cohorts.Indeed, millennials have been exposed to specific macro changes that may have impacted their paths to residential independence and the role played by certain labour market conditions in the risk of leaving parental home for different reasons.
The economic crises of the new century-the 2008 Great Recession and the 2010 current sovereign debt crisis-hit young people particularly hard, hindering their efforts to achieve financial independence.During the recessions, youth unemployment rose disproportionately with respect to the overall unemployment level in almost all European countries (Aassve et al., 2013a(Aassve et al., , 2013b;;Cho & Newhouse, 2013;Lin et al., 2013;OECD, 2013) and temporary or informal contracts became more common.This made young people extremely financially vulnerable (Marcus & Gavrilovic, 2010) as such contracts were easily at risk of ending during these economic cycles.The conjunctural economic crises of 2008 and 2010 might therefore have further amplified the economic vulnerability of younger cohorts, inducing an additional delay in young Italians' first residential autonomy (Carcillo et al., 2015).Recent ISTAT data suggest that this could be the case: the percentage of young people (aged 18-34) citing economic reasons as their main motive for staying in the parental home has increased over the years. 3Alternatively, the generalized climate of economic uncertainty experienced by millennials may have made them more accustomed to insecurity, meaning that even in the case of precarious employment, the new generations might be more willing to leave the parental home for less long-term commitment reasons (i.e.non-union-related reasons or cohabitation).
Millennials have also been exposed to other macro changes of a behavioural and cultural nature, potentially impacting their paths to residential independence.They are, for example, the first to have fully experienced a European 'youth culture' , aspiring to individual rights, post-materialistic values and gender equity, whose spread has been significantly accelerated by the pervasiveness of internet connections and young adults' increasing international mobility (Cepa & Furstenberg, 2021;Digital, 2022;Leccardi & Ruspini, 2006).A rising rate of tertiary education over the cohorts (especially among women) has indubitably also supported this process.This has been accompanied by a rising age at completion of education (ISTAT, 2014b).All these cultural changes may have influenced the lifestyles and preferences of youth, including the conditions affecting their risk of leaving the parental home for different reasons.On the one hand, they may have changed millennials' preferences, making them more likely to exit for non-union related reasons even in instances of mild economic vulnerability (i.e.job precariousness).On the other hand, increasing gender equity may have attenuated the traditional male-breadwinner model, making the impact of economic vulnerability on departure for union reasons less gender differentiated.Moreover, millennials are the first generation to fully experience increasingly widespread 'modern' family behaviours.Though the SDT started later in Italy than in other Western and Northern European countries, the first 15 years of the new century witnessed a sharp rise in cohabitation, out-of-wedlock childbearing and divorce (ISTAT, 2015(ISTAT, , 2017(ISTAT, and 2022)), suggesting that the family behaviours of Italians, especially of the more recent cohorts, are becoming more similar to those of their European counterparts (Pirani & Vignoli, 2016).The spread of non-marital unions in particular may have impacted the paths millennials' take out of the parental home and, with this, even the role played by job uncertainty in departures for union-related reasons (Vignoli et al., 2016).
Given the changes involving the most recent cohorts, we formulate the following hypotheses with the aim of verifying whether the role of labour market statuses in departure from the parental home has shifted over the birth cohorts: HP5: Labour market exclusion has a lower gender specific impact on exit for unionrelated reasons among the most recent cohorts (millennials) compared to previous cohorts.The increasing gender equity observed in Italian society in the last decades leads us to expect that the effect for women has become more similar to that for men.Meanwhile, the effect for men is expected to remain negative, as they continue to be the main economic supporters of the family.HP6: Precarious employment (temporary contracts and self-employment) has a less negative effect on the risk of exit for union-related reasons among more recent cohorts than among the previous cohorts, both for men and women.The spread of consensual unions over the last two decades leads us to expect that more recent generations may exit in greater proportions for cohabitation than for marriage compared to previous cohorts.As cohabitation is a less binding union than marriage, we expect that the negative role of precarious employment is attenuated for men and women of more recent generations.HP7: The potential negative effect of precarious employment (temporary contracts and self-employment) on the risk of exit for non-union-related reasons is weaker among more recent cohorts for both men and women.Since precarious employment has become more and more widespread over the generations (temporary employment, for instance, became the dominant way of entering the labour market for young people), the perception of economic insecurity associated with it has arguably attenuated.We therefore expect the decision to exit the parental home to depend less on job type than in the past.

Data, selection of the sample, preliminary descriptions, and statistical implementation
recorded for individuals who at the time of interview had left the parental home.This means we can study not only whether young adults have left the parental home, but also, in the affirmative, the pathways taken (marriage, cohabitation, work, study, independence, death of a parent, other reasons).
We focus on young adults who at the time of interview (in 2009 and in 2016, respectively) were between the ages of 20 and 44 (and thus, born between 1965 and 1996).Following previous studies (e.g.Blaauboer & Mulder, 2010;Schwanitz et al., 2017), we use an age range of 16 to 34, as leaving home before or after these ages is generally considered to be out of the ordinary, requiring more in-depth study.We similarly exclude cases where the respondent reported having left the parental home for the first time before the age of 16, seen as either unrealistic or outliers (220 observations in 2009 and 150 in 2016, respectively, corresponding to 1.6% and 1.7% of young adults ages 20 to 44 at interview).We also drop cases with missing information on whether or not respondents have left the parental home (148 observations, 1.7%, only for 2016).Finally, cases are censored at the time of interview or at age 35 in instances where transition out of the parental home has not yet occurred.As a result, our sample consists of 22,218 respondents (50.5% women and 49.5% men), among whom 66.1% had left the parental home by the time of the interview (or by the age of 34, if the respondent is aged 35 or older at interview).
Table 1 provides more detailed descriptive statistics, considering gender and cohort differences as well as specific reasons for departure.We observe notable differences between the sexes: lower percentages of young men have left the parental home compared to women (60.2% vs 71.8%), and while women have mainly exited to go live with a partner, young men have departed in the same proportions for union-and non-unionrelated reasons.Gender differences also emerge among departures for non-unionrelated reasons, with a higher proportion of young men leaving home for work motives compared to women, and greater proportions of young women exiting to study.
To test the hypotheses formulated above, we group individuals into two main categories: those born between 1965 and 1979 and those born between 1980 and 1996 (millennials).We also divide the reasons for leaving into two broad categories: unionrelated (marriage and cohabitation) and non-union related (work, study, independence and other).This allows us to obtain an adequate sample size by gender for each group defined by cohort and reason for exit (Table 2).

Leaving the parental home across the birth cohorts: preliminary descriptive analysis
Figure 1 presents the survival functions of leaving the parental home for men and women grouped into the two above-described cohorts: 1965-1979 and 1980-1996.Figure 2-which estimates the survival functions using a competing risks approachconsiders the reasons for leaving, distinguishing between union-related (Fig. 2A) and non-union-related (Fig. 2B)5 motives.
In the first figure, we see that there are almost no differences across cohorts in the timing of departure from the parental home, for both men and women.However, a very different picture emerges when considering the reasons for leaving in the second figure.
Younger cohorts leave the parental home for union-related reasons later than older cohorts (Fig. 2A), and this is true for both men and women.Indeed, survival functions are higher for younger cohorts, meaning that at every age, the percentage of individuals  In contrast, and especially for women, younger cohorts leave earlier for non-unionrelated reasons than their older counterparts (Fig. 2B).For example, while 73% of women born between 1965 and 1979 have not yet left the parental home at age 30 for non-union related reasons, this percentage drops to 61 for women born between 1980 and 1996.The difference between the two male cohorts is smaller: 67% for the older group and 63% for millennials.These results suggest that millennials, and particularly women, anticipate leaving home for non-union-related reasons and postpone leaving for union-related reasons.

The models
We use event history models to shed light on the process of leaving home among the cohorts of young Italians born between 1965 and 1996, paying particular attention to the role of economic vulnerability and possible changes in its impact-both on the timing and reasons for exit-across the different the cohorts.Specifically, since only the year of leaving the parental home is recorded, we estimate discrete-time event history models, adopting a person-year scheme.In addition, in order to answer our different research hypotheses, we conduct separate analyses for men and women (an approach commonly used in the literature; see, for example, Aassve et al., 2002;Blaauboer & Mulder, 2010;Chiuri & Del Boca, 2010;Iacovou, 2010;Modena & Rondinelli, 2016;Rusconi, 2004;Schwanitz et al., 2017;Sironi et al., 2015).Here, men are observed along 118,416 personyears and women along 107,586 person-years.
First, we consider the process of leaving home as a single 'destination' (thus estimating a logistic regression of person years) and then use a competing risks approach (multinomial logistic regression), where leaving the parental home to live without a partner and leaving home to live with a partner are the outcomes of interest (a strategy that follows previous studies; see, for example, Blaauboer & Mulder, 2010;Mulder & Clark, 2000;Zorlu & Mulder, 2010).As mentioned above, the age range of 16 to 34 is used and cases are censored at the time of interview or at age 35 if transition out of the parental home has not yet occurred.
The discrete-time logit model used in a single destination perspective (Allison, 2014) is given by: where h it is the probability of leaving the parental home at time t for the young adult i (who has not yet departed).
In using a competing risks approach, we can allow the determinants to differ between the decision to leave the parental home for non-union-related reasons and that for union-related reasons (Iacovou, 2010;Schwanitz et al., 2017).In this perspective, the transitions to different states are considered to be competing events, and the coefficients of the model are the effects of covariates on the probability of moving into a state i rather than remaining in the reference state.
The discrete-time multinomial logit model is given by: where h (r) it is the probability of leaving the parental home (state 0) for reason r at time t for the young adult i, and x itj (j = 1,…, k) is the covariate vector (with some time-varying covariates), with its corresponding coefficients β (r)  j for exit reason r.

Independent variables
In this framework, our core explanatory variables are individuals' birth cohort and labour market situation.All other covariates are used as controls.

a) Key covariates
The respondents' birth cohort is defined as a dichotomous covariate that distinguishes millennials (born between 1980 and 1996) from previous cohorts (1965)(1966)(1967)(1968)(1969)(1970)(1971)(1972)(1973)(1974)(1975)(1976)(1977)(1978)(1979). 6Labour market situation is a (time-varying7 ) variable that combines employment status and job type, thus capturing potential economic vulnerability in terms of either labour market exclusion (no personal income) or objective job insecurity (uncertain income).Indeed, more than simply whether the individual is or is not employed, we consider job stability using information on occupational status and type of contract.Specifically, we construct four categories: not employed individuals (including inactive and unemployed individuals8 ), fixed term employees (employed with a temporary contract), permanent employees (permanent contract), and self-employed individuals9 (young entrepreneurs and freelancers).
To properly isolate the cases of unemployed and inactive individuals from students, this variable is used in connection with school enrolment condition-a time-varying10 log h dichotomous variable that takes into account whether or not the individual is in education programme.11

b) Control covariates
Control covariates influencing timing and pathways out of the parental home can be differentiated between: individual, family and contextual characteristics (Schwanitz et al., 2017).
The individual characteristics that we control for in our analyses are age and level of education.As previous research consistently documents a strong age differentiation in leaving the parental home (Blaauboer & Mulder, 2010), we include age in the models in linear as well as in logistic form (different tests reveal that the linear-logistic specification shows the best fit [smallest AIC and BIC values]).Education level has, meanwhile, been linked to different living arrangement preferences and value sets (Mulder & Hooimeijer, 2002), where higher levels of schooling seem to be associated with non-traditional values and individual independence (Liefbroer & Billari, 2010).Here, we measure education using a time varying variable with three categories: high (university degreebachelor, master or PhD), medium (secondary school) or low (below secondary school).
Leaving the parental home is often the first event marking the transition to adulthood.The decision to depart is not, however, simply an autonomous choice of young adults but is also influenced by the context in which they grew up.We accordingly control for various family characteristics, including parental level of education, separation and maternal employment status.Parental education has, in fact, been shown to affect leaving home through the transmission of both cultural and economic resources (Blaauboer & Mulder, 2010;Sironi et al., 2015;Ward & Spitze, 2007), though findings are mixed and results depend both on the context and pathways of leaving home (Blaauboer & Mulder, 2010;Iacovou, 2010;Manacorda & Moretti, 2006;Schwanitz et al., 2017;Sironi & Rosina, 2015;Sironi et al., 2015).In our measure of parental education, we consider the education levels of both parents at the time of interview.Specifically, we define a unique set of categories that takes into account the highest education level obtained by at least one parent: 'high' refers to respondents who have at least one parent with a university degree; 'medium' where neither parent has a university degree, but at least one obtained a secondary school diploma; and 'low' includes the remaining individuals with low-educated parents (neither having completed secondary school).In addition to the importance of parental education, studies also highlight parental separation as a determinant of leaving home (Blaauboer & Mulder, 2010;Mazzuco & Ongaro, 2009;Tosi, 2017).Here, we use a time-varying perspective to consider whether or not respondents' parents are divorced or separated, as discerned from the survey records, which indicate the year in which the couple likely stopped living together.Finally, research suggests that mother's employment status may be a significant contributing factor in young adults' decision to leave home (Holdsworth, 2000;Iacovou, 2010).We accordingly include a dichotomous variable that distinguishes between employed and not employed mothers, derived from a question about the mother's job situation when the respondent was aged 14.
Lastly, we control for area of residence (at interview) as a contextual characteristic, allowing to account for geographical heterogeneity and the impact of unobservable context variables.Italy is characterized by deep territorial differences and, especially, a North-South divide, making it particularly important to consider this dimension.Young adults living in the North behave differently from those in the South, varying in their timing of departure and pathways taken out of the parental home (Bertolini et al., 2017;Sironi & Rosina, 2015).In the analyses that follow, we distinguish between four areas: Northeast, Northwest, Centre, South and Islands.
Table 3 presents descriptive statistics for the subset of time-fixed variables12 for men and women in our sample.

The determinants of leaving home
We begin, as mentioned, by considering the exit process as a single destination.Tables 4  and 5 report the results of the models estimated separately for, respectively, young men and women, where we make no distinction in terms of the reasons for leaving.
Model 1a reveals that, net of other confounding factors, male millennials have a significantly lower risk of leaving the parental home compared to men in previous cohorts.In contrast, there are no significant differences between female millennials and women of previous cohorts in this regard (Model 1b).Models 2a and 2b confirm this result.
As concerns the role of labour market status, models 1a and 1b show that not being employed lowers the likelihood of leaving the parental home for both young men and women, compared to permanently employed young adults.Meanwhile, a fixed term contract does not influence the risk of departure from the parental home (for both men and women-coefficients are small and not significant).A situation of relative precarity such as self-employment instead negatively impacts the risk of leaving the parental home for young men (though the coefficient is small and weakly significant) but not for young women (the coefficient is small and not significant).Model 2b confirms these results for women (the interaction terms are not significant), suggesting that the role of economic vulnerability has not changed over the different female cohorts.We observe, however, an entirely different dynamic for young male millennials (Model 2a).Specifically, the negative effect of labour market exclusion decreases due to the significant positive coefficient of the interaction term.Noteworthy as well is the fact that male millennials with fixed-term employment show a higher risk of departure from the parental home compared to their counterparts in previous cohorts.A situation of relative uncertainty such as self-employment instead has the same negative impact on the risk of leaving the parental home across cohorts.Thus, if for young women, whatever the cohort, only a situation of labour market exclusion decreases the likelihood of leaving home, for male millennials new patterns seem to emerge.For this group, being out of the labour market continues to have a negative effect on the risk of departure, albeit less strong compared to previous cohorts.Meanwhile, a situation of economic vulnerability such as a fixed-term contract increases the risk of leaving home compared to previous cohorts.These results should, however, be read with caution, as not distinguishing the reasons for exit might hide opposite behaviours connected with different employment conditions.

Exit for union-related and non-union-related reasons
In this section, we present the results of the competing risks models, again estimated separately for young men (Table 6) and women (Table 7), where leaving the parental home for union-related reasons and leaving home for non-union-related motives (to pursue education or employment opportunities or to establish independence) are the outcomes of interest.Accounting for this dimension is illuminating, revealing that the process of departing the parental home for union-related motives is notably different than doing so for other reasons.a) Exit for union-related reasons Models 1c and 1d show that both male and female millennials have a lower risk of leaving the parental home for union-related reasons compared to previous cohorts.This pattern is confirmed in models 2c and 2d.
Models 1c and 1d demonstrate that labour market exclusion, compared to permanent employment, lowers the likelihood of leaving the parental home to go live with a partner for young men (Model 1c), but not for women (Model 1d).Precarious employment conditions (i.e. a fixed-term contract) lower the risk of departure for union-related reasons for both men and women.Meanwhile, a situation of relative uncertainty, such as self-employment, does not affect either young men or women's risk of leaving the parental home to go live with a partner.Models 2c and 2d confirm all of these results for both sexes, implying that the role of economic vulnerability is the same among millennials and previous cohorts.
To summarize, economic vulnerability connected to labour market exclusion only has a negative effect on the risk of leaving the parental home for union-related reasons for men, not for women.This confirms our HP1, which is based on the persistence of the male breadwinner model in Italian society.Contrary, however, to HP5, the impact of being not employed has not changed for women across the cohorts, suggesting that changes in gender roles in Italy do not imply a convergence of the effect of labour market exclusion for women towards that for men.Meanwhile, the   negative effect of having a precarious job on the risk of leaving home with a partner is confirmed, but only for temporary contracts (for both men and women).Moreover, being self-employed is less detrimental on leaving home for union-related reasons than a fixed-term job.Our HP3 is thus partially confirmed.The negative effect of temporary contracts is, however, the same across cohorts, thus rejecting HP6.It would thus seem that the spread of less binding forms of union, such as cohabitation, has not impacted the role played by precarious employment among millennials.
b) Exit for non-union-related reasons Models 1c and 1d reveals that, net of the effect of other covariates, female millennials have a higher likelihood of leaving the parental home for purposes other than to live with a partner, while male millennials instead have a slightly significant lower risk of departure from the parental home for non-union-related reasons compared to previous cohorts. 13he role of economic uncertainty and its differing impact for millennials compared to previous cohorts on departure for non-union-related reasons notably contrasts with the findings for union-related exits detailed above.Labour market exclusion, but also uncertain work conditions as defined by self-employment, decrease the risk of leaving the parental home without a partner for both men and women (models 1c and 1d), though the effects, particularly those connected with self-employment, are less strong for male millennials compared to previous cohorts (model 2c).A situation of uncertainty as defined by a fixed-term contract instead increases the risk of leaving home without a partner (particularly) for men, and this effect is stronger for male millennials.The interaction terms are not significant for women, indicating that the effect of economic vulnerability is the same among female millennials and their counterparts in previous cohorts.
As assumed in HP2, labour market exclusion is thus detrimental for leaving the parental home for non-union-related reasons for both men and women.Interestingly, the effect has not changed across cohorts.Thus, having a minimum of personal income is a pre-requisite for autonomy, in a country where the welfare system does not support individuals in obtaining personal independence.The impact of having a precarious job differs depending on whether this consists of fixed-term employment or self-employment.Self-employment decreases the propensity to leave the parental home without a partner for both men and for women, as expected in HP 4, and this effect diminishes among the recent cohorts in the direction supposed by HP 7, though only for men.When, however, it comes to temporary contracts, HP4 is rejected.Contrary to our expectations, economic insecurity connected with fixed-term employment does not delay leaving home for non-union related reasons.Rather, our results suggest that both men and women with a temporary contract show a higher risk of leaving home compared to individuals who are permanently employed.In addition, in the direction of HP 7, male millennials with temporary employment have an even higher propensity to leave the parental home for non-union-related reasons compared to their counterparts with permanent jobs.

c) Control variable results
The outcomes for the control variables are, generally, unsurprising (Tables 6 and 7).As expected, being a student decreases the likelihood of leaving the parental home to form a union, and increases the risk of leaving for non-union-related reasons for both men and women, irrespective of birth cohort.Men and women with lower levels of education have a lesser risk of leaving the parental home for non-union-related reasons and, instead, a higher risk of leaving to form a union.Higher educational levels may be associated with more non-traditional values and a desire for independence, thus increasing the risk of leaving the parental home for non-union-related reasons (Liefbroer & Billari, 2010).
Our findings on the impact of family background also go in the expected direction.Both men and women with higher-educated parents have a greater risk of departure for non-union-related reasons and a lower risk of departure for union-related reasons (Blaauboer & Mulder, 2010).As also observed by Mazzuco and Ongaro (2009), having separated parents and/or employed mothers accelerate women's departure from home for non-union-related reasons.The same result holds for men with employed mothers.Lastly, young men and women living in the Northeast have the highest likelihood of leaving the parental home, independently of the pathway.

Conclusion and discussion
Leaving the parental home is a key event in the transition to adulthood and a rich literature explores the determinants of this transition.In this study, we investigate how young adults' economic self-sufficiency-proxied by individual labour market status-affects the decision to leave the parental home.We furthermore distinguish between union and non-union related reasons for departure.The focus is on Italy, a country that in recent decades has experienced important changes both in labour market conditions and value orientations.We accordingly adopt a cohort perspective, allowing to assess whether the impact of labour market conditions on leaving home has changed over the generations.
Our results highlight the importance of considering the reasons for leaving the parental home.Indeed, making this distinction reveals that potential economic vulnerability associated with certain individual labour market positions has different effects according to the motives for exit.
Specifically, labour market exclusion has a strong negative effect on the exit for unionrelated reasons only for men, and its impact has not changed over the generations.The result is consistent with empirical literature suggesting that-in more conservative societies-inactive women may not be penalized in exiting the parental home for unionrelated reasons (Del Rey et al., 2022).Moreover, this finding indicates that Italy continues to be a conservative country, with a male breadwinner model persisting to some degree, even among the most recent birth cohorts.
In line with previous theoretical frameworks (see, for example, Mills et al., 2005), having a precarious job in the form of a temporary contract also has a negative effect on the risk of exiting the parental home for union-related reasons-for both men and womenand this effect persists across cohorts.Thus, millennials' seemingly greater ease with uncertainty and the recent spread of cohabitation (a less binding form of union than marriage) have not led to a change in behaviour compared to previous birth cohorts.Self-employment does not, however, seem to penalize men and women in exiting the parental home for a union.It may be that this type of work is less exposed to vulnerable economic conditions (level of income and future job prospects) than temporary employment.
Our findings furthermore show that not all labour market positions that are potentially vulnerable from an economic standpoint have a negative impact on leaving home for non-union-related reasons.On the one hand, labour market exclusion and uncertainty connected to self-employment do have a negative effect on the propensity to leave the parental home for both men and for women, though, as expected, these effects are less strong among male millennials.On the other hand, uncertainty connected to temporary contracts surprisingly, and contrary to our hypothesis, increases the propensity to leave the parental home for non-union related reasons.This is particularly true for men, additionally so among the more recent cohorts.This positive association is not easy to explain and a lack of data on the specific circumstances characterizing young adults' leaving home with temporary jobs means we can only speculate about possible mechanisms.
The hypothesis that this result might depend on differences in wages between temporary and permanent employment contracts (van Wijk et al., 2021), where the former are more economically advantageous than the latter, is questionable.On the one hand, temporary jobs in Italy are generally less well paid and offer poorer conditions than permanent ones (Istat, 2021;Picchio, 2006).On the other hand, further exploratory analyses on our data (not shown here) show that the individuals with fixed-term contracts have higher propension for leaving the parental home for non-union-related reasons than permanent employees, only if they are low educated.It might be that individuals with permanent employment prefer to delay departure in order to accumulate economic resources for more 'stable' household transitions (for example, becoming a homeowner) in comparison with those with a precarious job.More generally, young adults (particularly men) in uncertain economic positions may have different life-course prospects than those with a permanent job.In this perspective, the former (willing to leave home) might see this as (the only) opportunity to leave the parental home in a less 'committed' way.It is also possible that after a certain age, young adults wishing to obtain autonomy are a selected group willing to accept to leave parental home also with the risk of a temporary contract.
Whatever the explanations, this suggests that the choices of residential autonomy for non-union-related reasons are due to more complex mechanisms than those strictly linked to the perception of economic uncertainty associated with precarious job.In this perspective, it becomes relevant to better analyse the circumstances of leaving for non-union related reasons (level of income, future job prospects, and other parallel life course career perspectives) among individuals with a temporary job.This is important also to better understand the implications of this type of transition.Is this an intermediate and less 'committed' step that introduces (or accelerates) the transition to other more stable employment or family statuses?Or does it instead produce a sort of deadlock situation, where individuals find themselves unable to enter a union or parenthood?To what extent are young people who exit the home in this way at risk of returning to live with their parents?More detailed studies are needed to better understand the mechanisms underlying the interrelationship between exit from the parental home for non-union-related reasons and type of job, as well as the longer-term paths of young people who leave the parental home for such reasons, especially those who do so in uncertain economic conditions.
Finally, leaving the home for union-related reasons also begs further consideration, particularly research that distinguishes between departures for the purposes of marrying and cohabitating.This work could not do it for sample size limitations, but such explorations would help to better understand whether the diffusion of non-marital unions has been accompanied by a new pattern of exit from the family of origin, characterized by less commitment than marriage, and perhaps viewed as more compatible with economic uncertainty.

Fig. 2 (
Fig. 2 (Pseudo-) survival functions at leaving home for A union-related reasons and B non-union-related reasons by age and birth cohort

Fig. 4
Fig. 4 Proportions of young men and women who have not left the parental home for (A) union-related reasons and (A) non-union-related reasons by age and birth cohort

Table 1
Distribution of respondents according to the main reason for leaving home by time of interview or by age 34 (if respondent aged 35 or older at interview) according to gender and birthcohort † Truncated at ages 30-34 in 2009; ‡ truncated at ages 25-29 in 2009, and at ages 32-35 in 2016; *truncated at ages 20-24 in 2009, and at ages 27-31 in 2016; **truncated at ages 20-26 in 2016 In Italics: percentage conditional frequencies to those who have left with(out) a partner

Table 2
Absolute and percentage (in parentheses) frequencies in the sample by birth cohort and reasons for exit.Men and women Proportion of young men and women who have not yet left the parental home by age and birth cohort who have left the parental home for union-related reasons is greater among the older cohorts.For example, at age 30, 61% of men born between1965 and 1979have not yet left the parental home to live with a partner; this same percentage is 70 for millennials.

Table 3
Descriptive statistics(N = 22,218 observations, 226,002 person-years).Percentages on the whole sampleThe time-varying variables, related to labour market situation, enrolment status, educational level and parental separation, are not presented in this table

Table 4
Leaving the parental home: discrete-time event history model with a single destination.Estimated model coefficients on the probability of leaving home (models 1a and 2a).Men

Table 5
Leaving the parental home: discrete-time event history model with a single destination.Estimated model coefficients on the probability of leaving home (models 1b and 2b).Women

Table 6
Leaving the parental home for union-related or non-union-related reasons: discrete-time competing risks models (estimated model coefficients).Men

Table 7
Leaving the parental home for union-related or non-union-related reasons: discrete-time competing risks models (estimated model coefficients).Women