Family formation, parental background and young adults’ first entry into homeownership in Britain and Germany

Abstract Although previous research shows that family dynamics and parental socioeconomic status influence the timing of young adults’ first entry into homeownership, much less is known about how the role of family factors may vary across countries with different housing systems. In this article, we use panel survey data from Britain and Germany to compare how family life course careers and parental socioeconomic background influence young adults’ initial entry into homeownership in these two divergent national contexts. The results show that in Britain, first-time homeownership transitions are tightly synchronized with partnership formation. By contrast, in Germany first moves into homeownership typically occur later around or after the arrival of children. Parental owner-occupation accelerates entry into homeownership in both contexts, while the effects of other parental characteristics are relatively muted. Furthermore, the results highlight how individual socioeconomic factors are critical determinants of entering owner-occupation. This is particularly true in Britain where there is a strong socioeconomic gradient in first-time homeownership transitions.


Introduction
In many countries, entering homeownership for the first time is a major life course step that often has significant implications for disposable income, debt and wealth holdings, residential conditions and perceived social status. The timing of first-time homeownership is also a 'family affair' in two significant ways. First, a rich tradition of longitudinal research shows that family formation events (such as childbirth or partnership formation) alter residential preferences, perceived housing needs and resources in ways that influence the timing of first dwelling purchases (Dieleman & Everaers, 1994;Kendig, 1984;Mulder & Wagner, 1998). Indeed, in many Western societies the political promotion of homeownership in recent decades has been underpinned by a normative belief that owner-occupied housing provides the most suitable context for childrearing and family life (Ronald, 2008).
Second, public debates about whether housing systems are entrenching and deepening social inequality mean that scholars are also becoming increasingly interested in how parental background shapes entry into homeownership (Coulter, 2018;Lersch & Luijkx, 2015;McKee, 2012). Concerns have been voiced that macrotrends such as the declining availability of secure well-paid work, greater indebtedness (often to fund higher education), reduced public welfare provision, and enhanced problems of housing access and affordability are all making it harder for young people to muster the resources to become homeowners (Andrew, 2012;Dewilde et al., 2018;Lennartz et al., 2016;Lersch & Dewilde, 2015). This could make transitions into owner-occupation increasingly contingent on parental financial transfers or other downward intergenerational flows of economic and practical support (Heath, 2018;Mulder & Smits, 2013). Christophers (2018) argues that these practices reproduce inequality across the generations as access to homeownership becomes increasingly confined to children from more affluent backgrounds who have greater access to familial support (McKee, 2012).
In Britain, concern that access to homeownership is stratified and unattainable for many young families has led some commentators to cast Germany as a 'role model' example of a more equitable and hospitable housing system (Harper, 2016;Sanderson, 2017). The basic idea is that family formation events and parental background may be less relevant for young adults' tenure pathways in Germany because young people have weaker aspirations and incentives to enter homeownership as long-term renting is more affordable, secure, attractive and culturally accepted. By contrast, many young Britons believe that rapidly entering owner-occupation is essential for prosperity and housing security (Kemp & Kofner, 2010), although the growing financial difficulty of buying a first home is also acknowledged (McKee et al., 2017a). However, this may be an idealized comparative view as delayed first-time homeownership in Germany is also due to the constraints imposed by strict mortgage lending and high deposit requirements (Scanlon et al., 2014). These barriers could act to exclude young people from less advantaged families from owner-occupation, thereby deepening the intergenerational transmission of wealth inequality (Kurz, 2004).
As many previous studies of family trajectories and first-time homeownership only examine patterns in one selected country (Andrew, 2012;Ermisch & Halpin, 2004;Kurz, 2004;Mulder & Smits, 2013;€ Ost, 2012), this study takes a comparative approach and asks how are family formation events (primarily childbirth and partnership formation) and parental background associated with young adults' first entry into homeownership in Britain and Germany? To answer this question we use rich panel survey data gathered over a long timeframe by the German Socio-Economic Panel (SOEP) and the British Household Panel Survey (BHPS)/United Kingdom Household Longitudinal Study (UKHLS). As these are ongoing studies, we are able to extend previous cross-national research conducted using data gathered from older cohorts who entered housing systems under very different constellations of contextual conditions (e.g. Andrew et al., 2006;Clark et al., 1997;Mulder & Wagner, 1998).
The rest of the article is structured as follows. The next section presents a theoretical framework to explain how individual and household factors, parental background and contextual conditions intersect to influence the timing of first-time homeownership transitions in Britain and Germany. We then describe the research design in the third section, before presenting and interpreting the results in the fourth. The final section concludes by discussing the broader significance and implications of our findings.

Theoretical framework
Conceptual models of housing behaviour have long posited a close link between family processes and housing careers. In the post-war decades, Anglophone life cycle models framed the transition to homeownership as a normative event that usually took place in early adulthood when family formation created a preference for the residential stability, leafy suburban neighbourhoods, enhanced social status and spacious dwellings provided by owner-occupation (Clark & Dieleman, 1996). These models recognized that family socioeconomic status stratified access to homeownership as poorer households and those from less advantaged backgrounds lacked the resources to purchase housing (Kendig, 1984). Although scholars have since cast aside the more rigid aspects of life cycle theory, researchers continue to draw on the basic notion that entry into homeownership is influenced by two sets of factors: the perceived relative costs and benefits of owning and one's ability to overcome resource barriers (Andrew et al., 2006;Mulder & Wagner, 1998;Mulder & Smits, 1999). Both of these decision-making components are themselves influenced by family life course dynamics, intergenerationally 'linked lives' and the balance of housing opportunities and constraints that are generated inter alia by national differences in tenure structures, housing market conditions, housing policies, mortgage markets and cultural attitudes towards homeownership (Kemeny, 2006).

Family dynamics and life course careers
Previous research shows that family dynamics and the structure of life course careers are influential predictors of first transitions into homeownership (Clark et al., 1997). Partnership is a particularly crucial factor because forming a co-residential relationship has strong impacts on residential preferences and also often improves access to resources. 1 According to Feijten (2005), becoming a homeowner is a major financial and locational commitment because the transaction costs of relocating are much higher for owners than renters. Single individuals who expect to form a co-residential partnership and who expect to be mobile for their education or career development, therefore, often defer homeownership until they form a settled relationship. For many people (particularly women), forming a partnership usually increases household income and access to the resources required to purchase a dwelling. This probably explains why dual earner couples are considerably more likely to become homeowners than single-earner couples (Clark et al., 1997). Importantly, in some contexts partnership type also shapes patterns of homeownership. In Germany and some Southern European countries it is marriage that boosts homeownership; while in Sweden, Great Britain and the Netherlands there are few differences between the effects of cohabitation and marriage (Lauster & Fransson, 2006;Thomas & Mulder, 2016).
Fertility events are also associated with homeownership transitions in some contexts. In liberal homeownership societies like the UK, limited access to secure, spacious, high quality rental accommodation in desirable locations for childrearing means that many couples enter homeownership as an indirect means to satisfy their new dwelling and neighbourhood preferences around the time of family formation (Mulder, 2006;Mulder & Wagner, 1998). In these contexts would-be parents also often strive to become homeowners because ideology, culture, mortgage marketing and the language of popular discourse 2 and policymaking all stress that owner-occupancy is a hallmark of adulthood, responsible citizenship and being a 'good parent' (Lauster, 2010;McKee et al., 2017a). By contrast, in countries like Germany with large unitary rental systems and a weaker ideological attachment to property ownership, greater access to rental accommodation adequate for families probably reduces the motivation to try and synchronize homeownership with family formation (Scanlon et al., 2014).
For poorer households, the links between fertility events and homeownership can be disrupted by limited access to resources. As childrearing is expensive and many new mothers reduce their working hours to provide childcare, having children could potentially 'crowd out' the resources that parents are able to devote to housing (Flynn, 2017). This may explain the aforementioned tendency for couples to try to enter homeownership before childbirth and thus potentially delay childbearing when homeownership is relatively less affordable (McKee et al., 2017a). Indeed, recent evidence suggests that these anticipatory homeownership transitions are waning in Britain as high house prices and credit constraints coupled with limited public welfare provision force a growing proportion of less affluent young people to form families in the private rented sector (Berrington & Stone, 2017).
Previous research indicates that several other life course factors influence housing preferences, resource access and thus whether people enter homeownership. Most studies show that first-time homeownership transitions accelerate as young adulthood progresses and people accumulate resources and complete the life events (e.g. finishing education and finding a long-term partner) that cultural scripts emphasize should occur before homeownership (Andrew, 2012;Mulder, 2003). Age may be particularly relevant in countries like Germany with more regulated or restrictive lending practices where people have to muster more savings to put down larger deposits (Blaauboer, 2010). In some contexts, greater difficulties and possibly discrimination in access to mortgage finance may explain why ethnic minorities and migrants frequently have lower rates of homeownership than the native majority (Davidov & Weick, 2011).
As mortgage lenders typically require deposits and the ability to meet loan repayments, in all contexts people with secure jobs and higher incomes are more likely to both prefer and have the means to enter homeownership than those with lower wages and little job security (Clark et al., 1997). Indeed, some authors argue that anaemic income growth and the shift to flexible, part-time and insecure labour markets has caused the recent fall in homeownership rates observed in some Western countries (Andrew, 2012;Arundel & Doling, 2017;Dotti Sani & Acciai, 2018). The social stratification of homeownership could also be more pronounced in countries with a marked insider-outsider division in the labour force (such as Germany or Southern European countries) than in countries with flexible labour markets (like the UK) where moving between jobs and occupational sectors is relatively easy (Lersch & Dewilde, 2015). In Germany, strong labour market regulations protect 'insiders' and leave 'outsiders' to face high levels of job insecurity (Lersch & Dewilde, 2015;Mills et al., 2005).
Educational qualifications improve access to homeownership in several ways. First, more highly qualified individuals tend to have higher and more stable incomes and thus greater access to resources. Second, tertiary qualifications signal to lenders that an individual is a less risky client with better job prospects and a lower likelihood of becoming unemployed. Finally, during higher education individuals may become socialized into more prosperous peer groups with a stronger ethos of homeownership and status competition. It is important to note that the positive effects of education on homeownership do, however, take some time to appear as graduates initially enter the labour market several years after their less qualified peers (Ermisch & Halpin, 2004). In the UK, the shift to a loan based system of student finance and restricted welfare provision mean that debt repayments are also reducing the income that recent graduate cohorts are able to devote to housing.

Parental background and intergenerationally 'linked lives'
Cross-national evidence shows that younger adults from advantaged backgrounds (often operationalized as parental education, social class or tenure) are typically more likely to enter homeownership than their less privileged peers (Coulter, 2018;Lersch & Luijkx, 2015;Mulder et al., 2015). Although this can be partly explained by strong intergenerational continuities in educational qualifications and occupational attainments, Mulder et al. (2015) posit that several additional mechanisms tie parental socioeconomic status to child homeownership. On the one hand, more affluent parents are better placed to provide direct assistance to help their children overcome the financial barriers to entering homeownership. This help can take many forms: loans or gifts to boost savings; advancing inheritance payments; acting as a mortgage guarantor; allowing children to live cheaply at home; or providing regular financial transfers and informal support to reduce young people's outgoings as they save up for a mortgage deposit (Druta & Ronald, 2017;Lennartz & Helbrecht, 2018;Mulder & Smits, 2013). In Britain, market research suggests that parents from advantaged social classes and those with greater savings may be more likely to give direct financial support to their children to help them become homeowners (Jessop & Humphrey, 2014). Interestingly this process is also evident in Germany as Lennartz & Helbert (2018) show that children with less affluent parents are less likely to seek housing-related familial financial support than those with more affluent parents.
Direct support practices are not the only way that parental socioeconomic advantage bolsters the likelihood that young people become homeowners. Helderman & Mulder (2007) argue that intergenerational tenure continuities arise partly because young people tend to live near their parents in similar types of housing markets. This means that on average, those growing up in homeownership are more likely to become homeowners simply because they will tend to stay in parts of the country with a greater supply of owned stock (Helderman & Mulder, 2007). This process is known to occur in parts of rural Germany where there is also a tradition of family assisted self-building (Kurz, 2004).
Furthermore, active or passive socialization of tenure preferences in early life could partly explain why children are more likely to enter owner-occupation if their parents own their dwelling (Lersch & Luijkx, 2015;Rowlands & Gurney, 2000). In some societies, parental homeowners might be more inclined to encourage their children into owner-occupation to help their offspring attain the wealth gains that they themselves have accrued during recent decades of house price growth (Soaita & Searle, 2016. Inheritance can also contribute to intergenerational continuities in tenure. This is particularly the case in countries with limited mortgage markets where prospective homeowners rely heavily on family housing support .

Divergent housing systems: the case of Britain and Germany
Long-running debates about the merits of classifying countries into 'housing regimes' (Kemeny, 2006) or clusters of institutional settings (Lersch & Dewilde, 2015) suggest that the national context shapes tenure preferences, as well as the housing opportunities and constraints that confront young people (Lennartz et al., 2016). This means that the influence of family dynamics and parental background on first-time homeownership probably varies across countries (Mulder et al., 2015). In Britain, concern for delayed access to homeownership and growing social inequalities in housing have led some commentators to argue that policymakers can learn lessons from Germany, which is thought to provide a more equitable, tenure neutral and sustainable housing system (Harper, 2016;Sanderson, 2017).
There are several reasons why preferences for homeownership differ in Britain and Germany. Perhaps the most regularly cited reason is that the cultural value attached to property ownership and the ideological promotion of owner-occupation is much stronger in the UK (Ronald, 2008). Since the 1950s, British governments of all colours have advocated homeownership, and, from the 1970s, introduced a raft of policy measures to help people become homeowners. These include the 1980 Housing Act's Right to Buy for council tenants through to the Help to Buy and Starter Home schemes introduced after 2010 by the Conservative-Liberal Democrat Coalition government. 3 At the same time, British fiscal policy has consistently favoured homeowners, who benefit from untaxed capital gains on owner-occupied housing and untaxed imputed rents. Until 2000 homeowners also enjoyed mortgage interest tax relief. Political promotion of homeownership has been especially pronounced in England, while Scotland has retained a stronger tradition and higher rate of social renting (McKee et al., 2017b).
This political promotion of homeownership has fuelled the ideologically framed popular British notion of a 'tenure hierarchy' where owner-occupation is seen as a mark of social success, while renting, particularly in the social sector, 4 is more stigmatized (Rowlands & Gurney, 2000). McKee et al.'s (2017a) interviews show that this discourse creates ambiguity, pressure and tension in young adults' housing preferences as many reproduce social norms by aspiring to homeownership while simultaneously recognizing that they are unlikely to ever be able to actually afford it. British cultural attitudes favouring homeownership are further bolstered by the popular lexicon of a 'housing ladder', as well as cultural output such as TV shows about home buying and home improvements (Gurney, 1999).
By contrast, in Germany the homeownership rate has been considerably lower than the Western average since the Second World War, most notably in the former East German regions (Zumbro, 2014). The German homeownership rate was 51.9% in 2015 as compared with 63.5% for Britain and the European average of 69.5% (Eurostat, 2015a(Eurostat, , 2015b. This has helped to prevent the development of an ideology or culture of homeownership in Germany, where it is still common to rent for a lifetime and renting long-term from a private landlord carries far less stigma than in Britain (Scanlon et al., 2014).
Differences between the British and German economic context and rental systems also shape housing preferences by influencing how people perceive the relative costs and benefits of ownership. In Britain, limited investment in new social housing since the 1980s combined with stock sales means that the sector has contracted dramatically (Murie, 2016). In many places, only the most vulnerable can now access more secure socially rented accommodation, while other renters are left to rely on a deregulated private sector where tenancies are short and rents can be high and increased at short notice (Kemp, 2015). 5 This means that many people seeking stable, long-term, high quality accommodation around family formation must strive for homeownership to obtain these housing attributes. High rates of house price inflation and price volatility further increase the relative attraction of ownership as the prospect of large wealth gains and fears about future price escalation combine to encourage people to buy dwellings (Soaita & Searle, 2016). Homeownership is also thought to function as both a supplement to scanty public pension provision and a good hedge against inflation (Doling & Ronald, 2010;Soaita & Searle, 2018).
Conditions are very different in Germany where there has been sustained private and (decreasing) public investment in rental housing since the 1950s (Voigtl€ ander, 2009). In Germany, tenants' rights are protected by indefinite contracts and rent caps, making private renting a more attractive alternative to homeownership that can meet people's housing preferences across a broader spectrum of life course stages (Easthope, 2014). Low rates of house price inflation until the early 2000s further reduced the motivation for Germans to try and make early transitions into owneroccupation (Voigtl€ ander, 2009), although in recent years prices have surged (Kholodilin & Michelsen, 2017). Some public financial support is provided towards mortgage costs through the pensions system, saving programmes supported by the government and subsidies targeting homeowners, particularly those with children (Davies et al., 2016;Mulder & Wagner, 1998).
Access to housing finance further influences the resources individuals in different countries can draw on to enter homeownership (Schwartz & Seabrooke, 2009). In Germany, conservative lending with high deposit requirements means that becoming an owner usually happens in midlife and is considered a 'once in a lifetime' event (Voigtl€ ander, 2014). There are a relatively restricted range of mortgage offers that are predominantly targeted to better off social groups (Davies et al., 2016). This is not true in the UK where a competitive mortgage market offering a wide range of products allows a broader cross-section of households to access housing finance (Toussaint & Elsinga, 2009). Although credit requirements became more onerous in the aftermath of the economic crisis, in recent years these restrictions have eased as taking on a large mortgage has once again become the only way for many young British households to surmount high house prices and enter homeownership.

Data and methods
This study uses three nationally representative household panel surveys conducting annual interviews with large samples of individuals. British data are drawn from BHPS (1991BHPS ( -2008 and its successor UKHLS (2009)(2010)(2011)(2012)(2013)(2014)(2015)(2016), also known as Understanding Society (University of Essex, Institute for Social and Economic Research, 2010Research, , 2016. BHPS began in 1991 when interviews were conducted with 10,300 individuals in 5500 households. Supplementary regional boost samples were added from Scotland and Wales in 1999 and Northern Ireland in 2001. In 2009, BHPS participants were invited to join UKHLS, a new and much larger panel survey with similar participant tracking procedures. UKHLS has a complex design consisting of a number of subsamples including a nationally representative General Population Sample as well as an Ethnic Minority Boost (EMB) oversample of key minority groups (Knies, 2017). In what follows we exclude EMB sample members as these have an unusually high attrition rate and are very concentrated in certain urban centres.
The German data are drawn from SOEP. SOEP began in 1984 with an interviewed sample of 12,245 individuals from 5921 West German households (Schupp, 2009). In 1990, SOEP was extended to include the former East German regions. General population and ethnic minority refreshment samples have been added periodically to keep the SOEP sample representative of the changing German population. For consistency with the UK analysis we do not draw on the SOEP immigrant refreshment sample.
The analytic sample consists of all fully interviewed, childless and single individuals living in the parental home at age 18 in any wave (n ¼ 4434 providing 15,757 person-years for Britain; n ¼ 6353 providing 38,063 person-years for Germany). 6 These cases are considered 'at risk' of moving into first-time homeownership between each consecutive pair of survey sweeps (t-1 and t). We follow these individuals until they move into an owned dwelling without their parents or until they do not participate in a survey wave. We do not allow people to re-enter the sample after a gap in their survey participation history because we cannot know about their housing experiences in the intervening period.
The dependent variable is coded as homeownership if an individual lives in an owned dwelling without a parent at t. We use this definition because SOEP does not provide complete annual information on which household members are the housing contract holders. For the analyses we distinguish three sets of independent variables: lagged individual characteristics, family-dynamics/life-events, and parental characteristics. The coding and measurement of these variables is harmonized as completely as possible across all three survey datasets.
The individual characteristics include an age variable centred on 18 (plus a squared term to capture nonlinear age effects) to act as the duration variable for the event history models. We also control for sex and foreign birth. As individual socioeconomic status strongly influences first-time homeownership we include variables measuring the education, income and employment status of sample members. Education is operationalized as a four-category 'highest educational attainment' variable: this distinguishes university degrees; non-tertiary vocational education; A-levels/GSCE/high school diploma; and lower-levels of education. These categories were chosen as they can be defined consistently for both countries. Income is operationalized as the timevarying total gross labour income of sample members plus their partner (if applicable). This variable is coded in three categories (low, middle, and high) with the cut points set to the 25th and 75th percentiles for that country-year. Employment status (time-varying) distinguishes five states: in secure employment (defined as working more than 30 h of work per week with a permanent contract); other employment; unemployment; full-time student and inactive. We also include period variable and lagged dummies to control for regional differences between Government Office Regions in Britain and the federal states (Bundesl€ ander) in Germany. 7 Due to space constraints the parameters of the period and region variables are not shown in the results tables (these are available from the first author on request). 8 Family dynamics variables measure leaving the parental home as well as family formation events that could occur between t-1 and t. These variables are defined as categorical indicators measuring the timing of the specified event. For partnership we use a variable with several categories: the year before entering a partnership; the year of partnership formation; the second year of the partnership; and the third or later years of the partnership. The reference category is remaining single. A categorical childbirth variable is operationalized in the same fashion. Categories within each of these variables are mutually exclusive and the variables reset and start recounting towards a new event at t-1. 9 The third set of independent variables capture information about parental family type (presence of biological and step-parents), educational attainments (university degrees), 10 tenure and number of siblings (including the individual). These variables are measured when sample members are first observed in the parental home at age 18. If parental information is not available in this wave then the variables are measured in the most recent year since the person was aged 14 (this is only necessary in the British data). To control for any imprecision this procedure may create we include a dummy to identify cases where parental information was not measured at age 18. Finally we include a dummy for urbanization. 11 The dependent variable is modelled using a series of event history analyses run separately for Britain and Germany. Separate models are used to allow the effects of each independent variable to differ flexibly across the two countries. As each individual contributes multiple observations to the models we estimated cluster-robust standard errors. We have also conducted several sensitivity checks to explore whether other differences between BHPS/UKHLS and SOEP might create a misleading picture of national patterns of first-time homeownership transition. As SOEP began in 1984 and we have no UK data prior to 1991 we have rerun the German models using only the data gathered after 1991. This did not alter the main findings. We have also refitted the models with a range of interactions between parental background and the period identifier, as well as interactions between parental factors and age. These extensions are discussed in more detail in the next section but in general have no impact on the main findings. The full results of these checks are available from the first author.

Results
Table 1 describes differences in young people's life courses and housing careers in Britain and Germany. Unsurprisingly, the first row of Table 1 indicates that during the study period first-time homeownership transitions were over 2.5 times more common in Britain (3.8% of observations) than Germany (1.4%). There are also clear cross-national differences in socioeconomic status. Reflecting institutional differences in the education system and labour market, the share of individuals with vocational school diplomas is substantially higher in Germany (15.4%) than Britain (8.2%), while less advantaged 'other' forms of employment (part-time or temporary contracts) are also much more prevalent in Germany (42.1% vs 19.9% in Britain). This is probably due to the relatively high share of temporary contracts in the German labour market (12.2% in Germany and 5% in Britain (Eurostat, 2015b)).
In contrast, the share of full-time students is much greater in the British (30.8%) than German (21.3%) data. This is probably partly because British individuals spend less time in the sample than their German counterparts as they tend to have faster transitions into first-time homeownership. The shorter time that young Britons are observed for probably also explains why they are less likely to have left home or be observed in a partnership than their German peers.
Importantly, there are also several cross-national differences in the parental variables which could impact on how parental socioeconomic advantage shapes children's first-time homeownership. The share of parents with a university degree is much higher in Germany (29.9%) than Britain (18.9%). While young Britons are far more likely to grow up in an owned dwelling than young Germans, the reverse is true for social housing (19.3% of British versus 6.8% of German parents). A larger proportion of young Britons also grow up in lone parent or reconstituted families, while the lower fertility rate in Germany (fluctuating between 1.25 and 1.56 in Germany as opposed to 1.63 and 1.92 in Britain between 1975 and 2010; Human Fertility Database, 2018) may explain why more German children grow up without siblings.
To examine whether these cross-national differences in life courses shape how family dynamics and parental background influence first-time homeownership transitions, Table 2 shows the results from two identical logistic regression models predicting initial entry into homeownership in Britain (columns two to four) and Germany (columns five to seven). 12 The variable effects are presented as log-odds coefficients and Average Marginal Effects (AMEs). AMEs indicate the population average change in the probability of entering homeownership for the first time between t-1 and t when the respective variable changes while all other variables are held at sample values (Williams, 2012). In both models, the likelihood of entering homeownership increases with age, albeit at a diminishing rate. The faster acceleration of entry into first-time homeownership through early adulthood in the British model confirms that young Britons have greater incentives and opportunity to become homeowners than their German counterparts. In both countries women are more likely to enter homeownership than men, although this effect is only statistically significant at the 5% level in Britain. These gender patterns could be due to women forming partnerships at younger ages with slightly older men. Interestingly, being in an urban area has a positive association with entering homeownership in Britain but a negative association in Germany. These divergent spatial patterns could reflect the affordability constraints imposed by the high costs of owner-occupied housing in the British countryside. By contrast, German urban areas are dominated by rental units which may limit younger adults' opportunities to buy housing in cities.
The models suggest that individual and household economics have a stronger effect on first-time homeownership transitions in Britain than Germany. Higher levels of education significantly accelerate first-time homeownership transitions in Britain, while education has no significant effects in Germany. Although higher incomes increase the relative likelihood of first-time homeownership transitions in both countries, the absolute effect of higher income shown by the AME is much higher in Britain (0.023) than Germany (0.010). The effect of employment status is also more potent in Britain where all people with all other economic statuses have a significantly lower likelihood of entering homeownership than their securely employed peers. By contrast, in Germany only unemployment significantly reduces the likelihood of firsttime homeownership entry compared to secure employment.
Taken together, these patterns suggest that entering first-time homeownership is a more common but also a more obviously stratified transition in Britain than Germany. This could be because most Britons strive to own but only those with resources are able to act on this desire. By contrast, in Germany reduced preferences and opportunities to own make this transition less common even among more advantaged social groups, thereby dampening housing tenure inequalities in early adulthood.
The effects of family dynamics vary across the two countries in ways which make it hard to generalize about the links between family and housing career development. In Britain but not Germany greater time spent outside the parental home significantly reduces the likelihood of entering homeownership. 13 This could be because relatively high rental costs in Britain make it hard to save for a deposit, while those renters living in social housing already have relatively secure accommodation and thus have fewer reasons to strive to own. While the results broadly confirm an association between family formation and first-time entry into homeownership in both countries, the AMEs signal much stronger synchronization with partnership than with fertility events. Synchronization with partnership is tighter in Britain than Germany, while fertility events only appear to be strongly related to first-time homeownership entries in Germany. Overall the results indicate that Germans tend to enter homeownership for the first time around or after family formation, whereas Britons enter earlier when forming partnerships. The insignificant effects of childbearing in Britain may reflect anticipatory moves into homeownership ahead of planned family formation. Alternatively, the lack of significant links between childbearing and first-time homeownership may be due to increasingly constrained access to homeownership around family formation during the study period (Berrington & Stone, 2017).
Intergenerationally 'linked lives' have a weaker association with first-time homeownership transitions in both countries than might be expected from public debates about family assisted housing purchases. Parental housing tenure appears to be the most potent familial factor in young adults' homeownership transitions Mulder et al., 2015). In both countries, children are much less likely to move into homeownership if their parents' dwelling is rented than if it is owned. This is probably due to tenure socialization and the children of renters having reduced access to family wealth. In both contexts the negative effect of renting is also more pronounced when parents rented from a social rather than private landlord. This result could indicate that parents living in social housing are also more economically disadvantaged in unobserved ways than their counterparts renting in the private sector.
Parental higher education has a significantly positive effect only in Britain where fewer members of the parental generation had university degrees than in Germany. 14 Having both biological parents at home at the age of 18 and the young person's number of siblings do not seem to have clear effects. 15 Overall, the models do not suggest that there is a far stronger intergenerational transmission of homeownership in Britain than in Germany. The absence of potent direct parental effects on child's first-time entry to homeownership also suggests that parental influence on children's housing careers may be primarily indirect and channelled through intergenerational continuities in education and occupational attainments.
To better evaluate the modelling results, Figures 1-4 use the models in Table 2 to compare the predicted probabilities that a hypothetical individual with a fixed set of characteristics enters homeownership for the first-time in Britain and Germany. 16 To facilitate comparison the y axes are fixed to have the same scale. Figure 1 shows the predicted probabilities of first-time homeownership entry around partnership events. In both countries, the probability of entering homeownership is greatest in the years around and immediately after partnership formation as compared with singlehood. In Britain first-time homeownership transitions often coincide with partnership formation events. Figure 2 shows the predicted probability of entering owner-occupation around fertility events. Here there are less clear-cut trends than in Figure 1. While child bearing is weakly positively associated with entry to first-time homeownership in Germany, in Britain the predicted probabilities are slightly lower around and after fertility events (with wide confidence intervals).
Turning finally to the parental effects, Figure 3 shows that children are somewhat more likely to enter homeownership in Britain if their parent(s) had a degree level education. By contrast, the effect of parental degree is negligible in Germany.      Figure 4 shows the effect of parental tenure. Children are most likely to become homeowners in both countries if their parents owned their own home, although the magnitude of this effect is small in Germany.

Conclusions
In many countries, reduced rates of homeownership in young adulthood have revived interest in the links between family processes and housing careers. In countries such as the UK, concerns have been voiced that enhanced difficulty accessing owner-occupation is decoupling family dynamics from housing careers by forcing less advantaged young people to form families in the private rental sector (Berrington & Stone, 2017). Across Europe, greater attention is also being directed towards the ways in which parental socioeconomic position influences child homeownership and thus could transmit (dis)advantage between generations (Christophers, 2018;Druta & Ronald, 2017;McKee, 2012). As research on family processes and housing often generalizes from Anglophone experiences or examines only one country at once, in this article, we harnessed comparable panel surveys to contrast how family formation and parental background shape first-time entry into homeownership in two very different housing systems. We concentrated on comparing Britain and Germany because their housing systems differ in multiple ways and UK debates about the housing challenges facing contemporary young people often hold Germany up as a 'good practice' model of a more equitable and sustainable housing system (Harper, 2016;Sanderson, 2017). Two conclusions can be drawn from the results. The first is that the links between family dynamics and housing careers differ across the two countries. In Britain, first-time transitions into homeownership were on average tightly synchronized with partnership formation in recent decades. In contrast, first-time homeownership tended to occur considerably later in Germany either around or after family formation. Staying longer outside the parental home in rented housing is also tied to a reduced propensity to enter first-time homeownership in Britain but not Germany. This could suggest that high private rental costs mean that young Britons benefit from remaining at home to save for a deposit (Druta & Ronald, 2017), whereas the timing of leaving the parental home is not so influential in the more affordable German private rental system.
Although transitions into first-time homeownership are more common for young Britons than their German peers, the results show that entering owner-occupation is somewhat paradoxically also a more socioeconomically stratified process in Britain than Germany. Higher education bolsters entry into ownership in Britain but not Germany and there is a stronger income gradient in the UK. Taken together, these results indicate that current British government initiatives to support young adults to enter homeownership are likely to disproportionately benefit more advantaged social groups. Given the magnitude of the structural trends placing downward pressure on young Britons' homeownership, the German results suggest that improving the cost and quality of rental alternatives to reduce the motivation to enter owner-occupation is a more efficient and equitable way to quickly improve young people's residential conditions. Our second conclusion is that there is a surprising degree of consistency in the association between parental socioeconomic position and child homeownership in the divergent British and German housing systems. Few parental factors other than housing tenure have any major independent links to first-time homeownership in either context. This could reflect measurement issues or alternatively may indicate that strong overall intergenerational continuities in housing are primarily a side effect of parental effects on other aspects of life course development, for example children's educational attainments or occupations (Ermisch & Halpin, 2004). The positive effect that parental owner-occupation has on children's transition to first-time homeownership is now well-documented across numerous countries and is probably a composite effect of wealth transmission, socialization and tenure acting as a proxy for unmeasured (dis)advantage (Mulder et al., 2015). Further work to unpack these mechanisms using innovative population level longitudinal modelling (Lersch & Luijkx, 2015) is now necessary, as are qualitative analyses of intergenerational practices of housing support that go beyond the middle class samples that have been the focus of most previous studies (Druta & Ronald, 2017;Heath, 2018). This will allow us to build up a richer picture of how parental background shapes housing careers and thus help design policies to support all young people enter and navigate housing systems. Notes 1. Homeownership entry can also occur alongside partnership formation when a nonowning partner moves in with an existing homeowner. Analysis of Dutch data indicates that this route is more common for women than men (Blaauboer, 2010 (Fitzpatrick & Watts, 2017). The generosity of housing benefit payments to renters has also been curtailed (Cole et al., 2016). 6. We use age 18 as our baseline as this is the minimum age one can independently buy a property in both countries. 7. Differences across devolved governments in Britain were initially tested by including dummies for the UK's component countries. However, models controlling for regional differences had higher predictive power and are thus reported here. 8. We use a linear period specification for parsimony as refitting the models with dummies showed a broadly linear pattern of diminishing transitions across time. 9. Checks suggest that using separate variables for cohabitation and marriage yield similar results. We have also modelled the effect of childbirth by parity but rejected this more complex specification because it did not show any systematic differences in the results.
Finally, a control for partnership dissolution was tested but then removed as it did not show any substantial effects. This is probably due to the very low number of dissolutions in the samples. 10. We originally added parental income but decided to drop it as it has no significant effects and did not improve the models. 11. Data constraints mean that urbanisation is operationalised in different ways in the British and German data. In Britain, urban areas are defined as settlements with a population of 10,000 or more. By contrast in Germany, 'urban' is defined as settlements with a population density of at least 150 people/km 2 or a city of at least 100,000 people with population density of at least 100 people/km 2 . 12. The pseudo r 2 values indicate that the models fit relatively well and have a similar level of explanatory power to those used in previous studies (Helderman & Mulder, 2007;Hubers et al., 2018). 13. Further testing showed that a linear term was the most parsimonious specification for this variable. 14. Additional models tested whether parental income and occupational status also predict first-time entry into homeownership. We found no evidence that these variables improved the model or were significantly associated with homeownership entry. This could indicate that parental income is a poor proxy for parental socio-economic status, or that it is other familial resources which matter for young people's housing career development. 15. We have also tested interacting parental factors with age and period. In Britain parental education and tenure seem to have slightly stronger effects on homeownership transitions as age increases, which may be because parents support their children in the housing market once they reach the normative age of living independently and settling down. However, these effects are quite small and thoroughly examining them is beyond the scope of this paper. 16. The profiles are set to have middle income, secure employment, and high school diploma, and be native-born, without a partner or child, middle-income parents with no degree, having both biological parents at parental home, having one more sibling, and living in Southeast England/Nordrhein-Westfalen.

Notes on Contributors
Sait Bayrakdar is a postdoctoral research fellow in the Department of Sociology at the University of Surrey. His research interests are social inequalities, education, migration, and housing.

Rory
Coulter is a lecturer in Human Geography at University College London. His research examines residential mobility and housing transitions over the life course.
Philipp Lersch is Associate Professor in Sociology of Social Policy at Humboldt University of Berlin and Senior Research Fellow at DIW Berlin. His main research interests are in social inequalities over and between life courses of individuals in divergent institutional contexts focusing on the family, wealth and employment.
Sergi Vidal is a Ramon y Cajal research fellow at the Centre for Demographic Studies -CERCA, Centres de Recerca de Catalunya. His research interests lie at the intersection between demographic trends (fertility, partnership, internal migration) and social inequality (housing tenure, labour force participation, union membership) from a substantive viewpoint, and on longitudinal and life course quantitative methods, from a methodological viewpoint.