Entering a Union in the Twenty-First Century: Cohabitation and ‘Living Apart Together’

Abstract

This chapter uses data from the HILDA survey to explore the relationship landscape of young Australians focussing on the transition to, and type of, first live-in relationship. Australians entered the twenty-first century having experienced 50 years of profound change in relationship formation, characterised by increasingly delayed marriage and a dramatic rise in unmarried cohabitation. Analyses of relationship formation behaviour almost exclusively focus on the live-in relationship. This chapter shows that social and economic factors result in delayed entry to cohabiting as well as married relationships. This delay has not led to a larger proportion of single people, but to the widespread uptake of LAT relationships in the early 20s. These LATs are considered serious by HILDA respondents, who indicate frequent contact and a high level of intention to cohabit with their LAT partners in the future. The chapter finds a strong age norm in relationship formation, a delay in entering unions, particularly marriage, and increasing selectivity of direct marriage in terms of education, religion and family background. It also finds that the relationship between direct marriage and education forms a u-shaped pattern with those with the least and the most education most likely to marry without cohabitation.

These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

2.1 Introduction

Australians entered the twenty-first century having experienced 50 years of profound change in the nature of relationship and family formation, occurring during a period of great social and economic transformation throughout the western world. This change has been characterised by a dramatic rise in cohabitation as an alternative to marriage and/or as a ‘trial’ before marriage. Much of the choice around relationship formation has been shaped by the increasing educational attainment and employment of women (Blossfeld and Huinink 1991). This chapter explores the early stages of adulthood, a time when young Australians are making decisions about entering into intimate relationships.

The most notable impact of this transformation on young people relates to increased access to reproductive control, changing attitudes towards sex and partnerships outside marriage and change in the structure of the labour market. The introduction of the contraceptive pill and easier access to abortion reduced the need for early marriage due to pregnancy. Alongside this were changes in attitudes leading to a widespread acceptance of non-marital relationships and a rise in the number of couples choosing cohabitation instead of marriage, particularly for first relationships (Evans 2013). The economy was also transforming with the modernisation and feminisation of the labour market. This lead to greater reliance on post-secondary education and a dramatic increase in post-secondary education for women.

This chapter describes the nature of first union formation in the first 10 years of the twenty-first century. It begins with a review of the literature on cohabitation in Australia to provide an historical context to current relationship formation patterns. It then provides an analysis of first relationship formation patterns of 18–30 year olds from 2001 to 2011, focussing on the choice between cohabitation and direct marriage.¹ Finally, it considers the nature and prevalence of ‘living apart together’ (LAT) relationships between 2005 and 2011.

The data for this chapter are drawn from the Household, Income and Labour Dynamics in Australia (HILDA) survey (see Technical Appendix). The data are used in three ways. Firstly, the data are used to examine the importance young people place on cohabitation and marriage. These data are drawn from a youth module collected in 2004. Secondly, data from 2001 to 2010 are used to model the choice of union type for first relationships. And finally, data from 2005 to 2008 are used to examine LAT relationships.

2.2 Cohabitation in Australia

2.2.1 Prevalence

The Australian Bureau of Statistics (ABS) first collected data specifically on non-marital cohabiting relationships in 1982. At this time it was estimated that approximately 5 % of all couples were living together without registering a marriage (Table 2.1). This rose to 16 % in 2011.

Table 2.1 The prevalence of cohabitation in Australia

Another measure of prevalence is the percentage of marriages in a given year that were preceded by cohabitation. Cohabitation prior to registered marriage has increased over the last 20 years. In 1992, just over half of all registered marriages were preceded by cohabitation (56 %) (ABS 1994). In 2012, over three quarters of marriages were preceded by cohabitation (78 %) (ABS 2013). The percentage of marriages preceded by cohabitation peaked at 79 % in 2010 and dropped slightly in the two following years.

These figures underestimate the experience of cohabitation in the population. While 12 % of couples were cohabiting in 2001, the percentage of people who had ever cohabited was much higher: using HILDA data, Dempsey and de Vaus (2004) estimate that 20 % of the ever-partnered population in 2001 had cohabited at some stage. Cohabitation is strongly associated with age. For those under 20 years of age who have ever been in a live-in relationship, 90 % have had at least one cohabiting relationship. This figure drops to 68 % at age 20–24 and 39 % at age 25–29 (Dempsey and de Vaus 2004).

These proportions all refer to people who have ever been in a live-in relationship. In the early stages of adulthood there are many people who have never been in a live-in relationship. To better gauge the spread of cohabitation across these younger age groups it is perhaps more important to consider those who have ever lived in a cohabiting relationship as a percentage of the total population. These calculations show that 18 % of 20–24 year olds and 29 % of 25–34 year olds have ever cohabited (Dempsey and de Vaus 2004).

2.2.2 Attitudes

In 1971 over two thirds (68 %) of married women living in Melbourne indicated that they would be “extremely horrified”, “considerably upset” or that they had “failed as a parent” if a son announced he was going to cohabit. If the announcement had been from a daughter, half (52 %) of these women would be more upset than if the announcement was from a son (Caldwell et al. 1988).

There have been various attempts to measure attitudes towards cohabitation in Australia. The Australian Survey of Social Attitudes in 2003 asked respondents to indicate whether they thought cohabiting couples, or cohabiting couples with children, constituted a family. There was general agreement that a cohabiting couple with children does constitute a family (79 %), but less so when children were not present (63 %) (Evans and Gray 2005).

The International Social Survey Program has collected data on attitudes to family over time. Questions were asked about cohabitation in 1994, 2002 and 2012. These data indicate that over the past 20 years attitudes towards cohabitation have become more liberal. In 1994, 28 % disagreed with the statement “It is alright for a couple to live together without intending to get married.” In 2002, 18 % disagreed and in 2012 only 14 % disagreed.

Using the 2005 HILDA survey, Qu and Weston (2008) find a higher (21 %) level of disagreement that “it is alright for an unmarried couple to live together even if they have no intention of marrying”. Among those aged 15–20 only 14 % disagreed and among those aged 20–29, 17 % disagreed.

2.2.3 Characteristics of Cohabiters

In the early 1990s, Glezer (1991) found that economic factors are important in the decision to cohabit, but love, friendship and companionship are also very important. Four studies since the turn of the twenty-first century touch on similar issues. Lindsay (2000) suggests that the reasons people choose cohabitation are more pragmatic than romantic. Lewis (2001) considers economic security to be a driver of cohabitation decisions. Cohabitation is seen as a rational response to low male wages. White (2003) finds that young people consider the 20s to be a period characterised by freedom and autonomy, and that there is some reluctance to partner seriously before age 30. Carmichael and Whittaker (2007) find that cohabitation ‘just happens’ as shared nights together increase.

In the 1980s Antill et al. (1983) reported the reasons men gave for cohabiting as having doubts about need for and nature of formal marriage, and greater gender equality in cohabiting relationships. Women agreed but also highlighted that they were not ready to settle down or to have children. They also found two major barriers to cohabitation: parental disapproval and religious objections (Antill et al. 1983).

Religion is strongly associated with relationship formation. Cohabitation is the highest among people who are not religious or claim no religious affiliation (Dempsey and de Vaus 2004; Glezer 1991). The largest religious groups in Australia (Catholic, Anglican) have the highest rates of cohabitation of all people who report a religious affiliation, whereas the lowest rates are found for those whose affiliation is Islam, Greek Orthodox, Sectarian or Pentecostal (Dempsey and de Vaus 2004). Khoo (1987) also finds cohabiters are less likely to be religious, as measured by patterns of church attendance.

Families and parents shape attitudes and behaviours surrounding cohabitation and relationship formation. Parental divorce is associated with cohabitation (Glezer 1991). People who have experienced parental divorce are more likely to cohabit than those who have never experienced divorce. Cohabitation is highest among Australians with English-speaking backgrounds (Khoo 1987; de Vaus 2004).

In the second half of the twentieth century cohabitation was seen as a relationship type of the educated middle classes. These were the ‘social trailblazers’ who experienced free tertiary education and high levels of economic opportunity. Glezer (1991) found that Australian cohabiters were most likely to have tertiary qualifications.

In the twenty-first century, research indicates that cohabitation, particularly as a setting for children, is now primarily a feature of couples with lower economic or educational status. A linear negative relationship with education has been termed the “pattern of disadvantage” and has been found across Europe and the US (Perelli-Harris et al. 2010; Gibson-Davis et al. 2005).

This pattern has also emerged in Australian research. Birrell et al. (2004) find that cohabitation is increasing for men with low education and income. Using census data from 1996 to 2006 Heard (2011) documents the change in Australian partnership patterns by level of education. The percentage of men and women cohabiting has increased at every age and for every education group, but rates of cohabitation are lowest among the tertiary-educated. For those under 30 the most dramatic increases have been among men and women with skilled vocational qualifications, resulting in an inverted U-shaped relationship between cohabitation and education: men and women with vocational qualifications have higher rates of cohabitation than do those with more or less education.

2.3 Importance of Cohabitation in Early Adulthood

The previous sections summarise our current understanding of cohabitation in Australia. But just how important is cohabitation to young people? In 2004 the HILDA survey included a youth module to collect information relevant to young people. Participants aged less than 30 were asked how important they felt it was for them to be living in a long-term relationship.² They were asked to consider this question across two time periods: How important it was at the time of the survey (‘now’), and how important they thought it would be for them at age 35.

The results show a dramatic difference between the present and future time periods (Fig. 2.1). For the present, the percentage indicating that living in a long-term relationship is not important decreases with increasing age. The percentage indicating that it is very important to them now increases with increasing age. There is virtually no variation by age in the percentage of young people who report that cohabiting is somewhat important to them now.

Fig. 2.1

The importance of cohabitation, now and at age 35

When asked about the future there is no difference in the responses from the different age groups even though there is a 15-year age spread. At each age 70 % of respondents predicted that it would be very important that they were living in a cohabiting union at age 35. It is interesting that there is no difference found here. Those aged 15–19 are looking forward 15–20 years, compared to 5–10 years for the 25–29 age group. It would be expected that those closest to the future age (35) would have similar responses regarding both the present and the future. This indicates a strong normative age for relationship formation among young Australians: most expect that forming a living together relationship would be a high priority between ages 30 and 35.

A similar question was asked about the importance of marriage but the pattern of responses was markedly different. For the present, respondents aged 15–19 placed similar importance on marriage as they did on cohabitation (Fig. 2.2). Among those aged 20–29 a larger percentage said that marriage was somewhat important than said that cohabitation was somewhat important. People in these age groups were more likely to say that marriage was not important than they were to say cohabitation was not important. The percentage who indicated that marriage is very important to them now did increase with age, as was the case for cohabitation, however only 21 % of those in their 20s placed a high importance on marriage.

Fig. 2.2

The importance of marriage, now and at age 35

The difference between now and age 35 is very similar for the 15–19 year olds irrespective of whether they are thinking about cohabitation or marriage. For the 25–29 year age group there is a different pattern evident when thinking about marriage in the future compared with cohabitation. The percentage indicating marriage is very important when considering the present increases with age. When considering the future the pattern is reversed, with older respondents placing less importance on marriage at age 35. As young people reach their late twenties, they place less importance on marriage occurring within the next few years.

The above results are based on never married respondents, but a significant proportion of young people are married before age 30. Figure 2.3 presents HILDA data on relationship status for each single year of age. Wave 8 (2008) of the survey is used as it is the most recent data with information on LATs. Below age 23, being single is the most common status and the combined proportions of respondents who are single and LAT sits above 50 % until age 24.

Fig. 2.3

Relationship status by age, 2008

Cohabitation grows very quickly in the early 20s and is dominant between ages 24 and 27. By age 28, marriage becomes the most dominant status and by age 30 close to 50 % of young people are married. This pattern of relationship status by age clearly reflects increased time spent in education and delayed relationship formation.

2.4 Entering the First Relationship

Young people in the early twenty-first century see living together relationships as an important aspect of their lives now and in the near future. More importance is placed on cohabitation than marriage. We also know that the majority of marriages are preceded by cohabitation. So what characteristics are associated with the choice of cohabitation or marriage for a young person’s first live-in relationship? This section uses prospective longitudinal data from HILDA to examine entry into the first live-in relationship. Live-in relationship (or live-in union) in this case refers to any relationship where the couple co-resides. This could be either a marriage or cohabitation. This analysis examines the impact of education, religion, and family background on whether the first live-in relationship is a marriage or cohabitation.

The sample includes everyone aged less than 30 years in 2001 (wave 1) who has not yet entered a live-in relationship. This group are followed across each wave of data collection until they experience their first relationship. The first relationship is identified as being either cohabitation or a marriage, which is the dependent variable for this analysis. Marriage here refers to direct marriage: that is, marriage where the couple have not lived together prior. As discussed earlier, it is expected that education, religion and family background are all important in determining the type of first live-in union. This is tested using a logit regression model and the results are presented in Table 2.2.

Table 2.2 Logit model predicting marriage over cohabitation for first live-in relationship

One of the difficulties in using education as a predictive variable for young people is that many of them may be still studying during the period of observation. In this analysis, education is measured at the wave where the first relationship is observed. The measure is the highest level of education obtained at that time: bachelor degree or higher, diploma or certificate, complete secondary or incomplete secondary. It has been shown that there have been changes in the composition of the cohabiting population and that cohabitation (for those under 30 years) is now most common among those with vocational qualifications and less common for those with no post-school qualifications or with bachelor degrees or higher (Heard 2011). However, it is not known if this same inverted u-shaped pattern is found for the type of first live-in relationship.

Previous research has consistently identified religion as having a strong negative association with cohabitation (Dempsey and de Vaus 2004; Glezer 1991; Khoo 1987). Various measures have been used to explore the impact of religion on relationship choice, including religious affiliation, church attendance, and importance of religion. The measure used here relates to the importance of religion in a person’s life. The categories are: important or very important, somewhat important, and not important. Fifteen per cent of the sample has missing information on this question so a missing category is also included in the model. The individuals with missing data are not significantly different from the reference category (somewhat important) in the models.

Families provide role modelling and values that are crucial to the development of individual attitudes and desires around relationship formation. Three measures describing the family background are used. The first is the ethnic composition of the family. The only measure possible for this analysis is whether one or both parents were born in a non-English speaking country. This measure is often used to determine differences that might be based on language or migrant status. Typically, those with overseas-born English-speaking parents behave in a similar way to native-born Australians. This measure has limitations as it includes migrants and non-migrants, those who have been here since infancy and those who have arrived recently, and those whose parents have been here since infancy or have arrived recently. These factors may be important since the environment in which an individual attended school, particularly secondary school, can have a large bearing on his or her attitudes to family formation.

The second measure of family background is whether or not the young person’s parents have been divorced. This measure includes those whose biological parents ever separated or divorced as well as a small number whose parents never married. Experiencing parental divorce can impact family formation decisions in both directions. Young people may be deterred from marriage as they see that it does not always work, or they may be drawn to marriage as a way to capture the intimacy and closeness that their parents did not maintain. Some people were not asked about parental divorce if one or both of their parents died before they were teenagers. An additional category is used to control for those cases where parental divorce is not applicable.

The third measure of family background is an attempt to capture the socio-economic position of the family. This could potentially be measured by parental income (only available for some of the respondents in the analysis), mother’s or father’s occupation, or mother’s or father’s level of education. Education is often used as a proxy for occupation or income. Given the nature of the data available in HILDA, and the overarching interest of this chapter in education, this analysis uses father’s highest level of education as an indicator of socio-economic status. The measure compares those whose fathers had no post-school qualification with those whose fathers did have a post-school qualification. An indicator is used to account for those who did not know their fathers’ highest level of education. The individuals with missing data are not significantly different from the reference category (no post-school qualification) in the models.

The results of the logit analysis, controlling for sex and age at first live-in relationship, are presented in Table 2.2. Education, as predicted, shows a u-shaped relationship with relationship type. Young people with a bachelor degree or higher have higher odds of the first relationship being a marriage rather than a cohabitation when compared to those who have completed secondary school. Those with incomplete secondary school also have higher odds of marriage than do those who completed secondary school. Interestingly this u-shaped relationship does not exist between education and first union type if the other variables in the model are not controlled for. Figure 2.4 presents the predicted probability of direct marriage by education level bivariately and controlling for the variables in the model reported in Table 2.2.

Fig. 2.4

Predicted probability of first live-in union being marriage by level of education

Considering education on its own (light bars), the predicted probability of direct marriage is 0.18 for those with a university degree. This stands out from those with lower levels of education where the predicted probability ranges between 0.06 and 0.08. When compared with the predicted probability of direct marriage from the model (controlling for religion, ethnicity, and family background) the nature of the relationship between marriage and education changes. Here (dark bars) the u-shaped relationship is obvious, as those with the least education and those with the most education both have a predicted probability of direct marriage of 0.14. This is double the probability of marriage for those who have completed secondary or a vocational qualification. This example highlights the importance of controlling for other factors in order to avoid spurious bivariate results.

Religion does not show a different pattern of prediction when controlling for the other variables in the model. It does, however, show a strong linear relationship with union type (Table 2.2). The more importance a young person places on religion in his or her life, the greater the odds of direct marriage. Compared to people who report religion being somewhat important in their lives, those who report religion being unimportant have lower odds of direct marriage. Those who report religion being important or very important have higher odds of direct marriage than do those who report religion being somewhat important in their lives.

The variables measuring the impact of family background also show an association with the type of first union. Parental divorce is associated with decreased odds of direct marriage. This suggests that, overall, parental divorce acts to deter young people from direct marriage in their first relationship. Having a parent born in a non-English speaking country is associated with increased odds of direct marriage when compared to those with both parents born in Australia or another English-speaking country. Father’s level of education is also associated with first union choice. The odds of direct marriage are higher for young people whose fathers have post-secondary qualifications than are the odds of direct marriage for those whose father have no post-school qualifications.

This analysis shows that education, religion and family background are all important predictors of direct marriage for first live-in union for young people in Australia. The results indicate that in an era with very high levels of cohabitation, direct marriage is becoming increasingly selective. Those who indicate that religion is important or very important are much more likely to choose marriage as a first live-in relationship. Direct marriage is also more likely for those with a bachelor degree or higher and those with incomplete secondary school.

2.5 LATs: An Alternative to Cohabitation or Simply Dating?

Very little is known about LAT relationships in Australia. Previous research is limited to a descriptive typology (Reimondos et al. 2011) and recent work on older people in LAT relationships (Malta and Farquharson 2012; Upton-Davis 2012, 2013). In 2005, the HILDA survey contained questions specifically related to LAT relationships. Reimondos et al. (2011), using this data, find that LAT relationships are widespread in Australia with 24 % of the ‘single’ population reporting a LAT relationship. LATs are more likely to be childless and never-married compared to single, cohabiting or married people.

The HILDA survey provides a unique opportunity to measure the incidence of LATs as well as other characteristics of these relationships. There has been debate in the literature about whether LATS are simply casual dating relationships or if they are a form of committed relationship that is being used as an alternative to cohabitation (Haskey and Lewis 2006; Ermisch and Siedler 2008; Trost 1998). For the age group being considered here (18–30 years) there are social as well as economic factors that would affect this distinction. The most obvious of these are education and the cost of setting up a home. Young Australians are leaving home at increasing ages partly due to increasing periods in education (Evans 2013; de Vaus 2004). This means that there is a longer period of time where LAT relationships are the most convenient. The cost of setting up a home and the low availability of rental properties in some areas of Australia could also influence young couple’s decisions to live apart. It is increasingly difficult for young people with lower and often part-time incomes to access the rental housing market.

This section will consider the frequency of LAT contact, cohabiting intentions, and transition to cohabitation for LATs aged 18–30. The data are again drawn from HILDA, specifically the 2005 and 2008 waves which contained questions about ‘intimate and ongoing’ relationships where the partners were not living together. For the data presented below a pooled sample of respondents in 2005 and 2008 is used to increase the number of respondents and increase the reliability of the patterns discovered. It has already been established that LATs are a significant relationship choice for young people, particularly up to age 23. Around 30 % of 18–20 year olds and 25 % of those aged 21–23 report being in a LAT relationship, dropping to a constant 10 % by the late 20s (Fig. 2.3).

2.5.1 How Often Do LAT Partners See Each Other?

LAT respondents were asked how often they saw their partners. In the two younger age groups, <20 and 20–24, 80 % of respondents saw their partners more than three times per week (Fig. 2.5). For those in the normative tertiary education age group (18–22), there is a greater tendency to meet almost every day. At this age there is a high level of homogeneity in the activities of young people, with partners moving and socialising in the same circles. It would be common for both partners to be in education and attending the same educational institutions. These factors make frequent contact relatively easy in the course of normal daily activities. As people get older and move into full-time working ages there is greater heterogeneity in their day-to-day lives. This is evident in the decrease in daily contact and the concurrent increase in less-than-weekly contact between partners. The category ‘3–5 days a week’ still remains the most common and 70 % see their partners more than 3 days per week.

Fig. 2.5

Frequency of meeting LAT partner by age group

2.5.2 Do LAT Partners Intend to Live Together in the Future?

The high level of contact between LAT partners does not reveal much about the seriousness of the relationship, how committed the partners are to each other, or how they see the future of the relationship. A better measure of whether couples are in LATs as an alternative to cohabitation is to ask about what plans individuals have for the future of their relationship. If LATs were considered as alternatives to live-in relationships we would expect there to be few people with plans to cohabit. This can also be a measure of the seriousness of the relationship. LATs are a part of the process of relationship formation and so for serious relationships it may be expected that people would be thinking about moving in together.

HILDA asked about the intentions of individuals in LAT relationships to cohabit with their current partners within the next 3 years. There was a steady increase across age groups: as people get older they are more likely to indicate that they intend to live together in the next 3 years (Fig. 2.6). A very high proportion (63 %) of those aged less than 20 indicated that they intended to live together over the next 3 years. This figure rose to 77 % for the 25–30 year olds. There is an element of social desirability in responses to this type of question (Lavrakas 2008) and it might be assumed that this is an overestimate of firm plans. Further, it is impossible to ascertain what the other individual in the partnership thinks, and whether he or she has similar or dissimilar intentions.

Fig. 2.6

LAT partners, intention to cohabit in the next 3 years by age group

2.5.3 Do LAT Partners Actually Move in Together?

Due to the way the data are collected it is not possible to determine if these intentions are realised. It is possible, however, to see if a respondent’s relationship status changed from ‘single’ to cohabiting or married in the year following the year he or she recorded being in a LAT relationship. In the wave that LAT status was recorded (2005 or 2008) the relationship status would have been recorded as single. HILDA provides the opportunity to look at the relationship status in the year following the LAT being recorded (2006 or 2009) to see if the respondent is still ‘single’ (maybe in a LAT relationship but maybe not – this cannot be determined) or if he or she is subsequently married or cohabiting. An assumption does need to be made that this marriage or cohabitation is with the same person as the LAT relationship in the previous year, as this is not measured. In a high proportion of cases this is probably true. If this assumption is held it is possible to estimate the proportion that move in with an LAT partner.

Figure 2.7 displays the distribution of relationship status in the year following the LAT relationship being recorded. At each age group the majority of these individuals are still ‘single’. In this case ‘single’ refers to all of those who are not in any relationship, those in a LAT relationship with the same person as the year before and those in a LAT relationship with a new partner. It is impossible to separate these groups using these data. While in the majority, the proportion ‘single’ does decrease with age. Increasing with age is the proportion cohabiting or married 1 year after recording an LAT.

Fig. 2.7

Relationship status 1 year after LAT identified, by age group

At the youngest age (<20 years), 12 % of LAT partners are in a live-in relationship 1 year later. For those aged 20–24 this figure is 19 % and 32 % for the 25–30 age group. It can be assumed that a reasonably high proportion of these relationships are with the same person as the LAT relationships recorded in the previous year. There is some lead time required for most relationships to progress to sharing a residence. A person would have to break up with an LAT partner, find someone else, enter a LAT relationship with that person and then move in with the new partner in less than 12 months for the live-in relationship to be with a different person from the LAT relationship. This is by no means beyond the realm of the possible, especially among the younger age groups, but would likely occur in a minority of cases.

So are LAT relationships a stage in the relationship formation process or being used as an alternative to cohabitation? For this age group (18–30 years) it is mostly the former. While half of respondents indicate that they have made a definite decision to live apart (analysis not shown), this is most likely related to their educational, financial and employment positions than a decision based on the relationship itself.

2.6 Discussion

This chapter sought to illuminate the factors that might affect the choice of a first live-in relationship type. It uses prospective panel data from a nationally representative source.

There is a strong age norm around live-in relationships. Living together in a cohabiting relationship is seen as very important between ages 30 and 35. When asked about the present, the importance young people placed on cohabitation increased with age. However, when asked about the future, young people across all age groups considered cohabitation by age 35 to be very important. There is an expectation that by the early 30s being in a cohabiting relationship will be a high priority.

During the 20s, other things are more important than living together. One of these is education. Changes in the labour market leading to the prolongation of education have affected the timing of relationship formation, leading to a delay in the onset of the first live-in relationship (Evans 2013). Analysis of relationship formation behaviour almost exclusively focuses on the live-in relationship. But this chapter shows that this delay has not led to a larger proportion of single people. This delay has instead led to widespread uptake of LAT relationships in the early 20s. These LATs are considered serious by the respondents who indicate a high level of intention to cohabit with their LAT partner in the future.

As education is prolonged, the timing of first live-in union is pushed later into the 20s. Cohabitation serves to provide a safe environment for young people to live together and be sexually active as if married, without the ties of marriage. This further delays marriage for most people who have one or more cohabitations prior to first marriage. Young people aged less than 30 place little importance on marriage in their present lives. They do, however, see marriage as having high importance in the future.

The inverted u-shaped pattern of relationship status by education found in this chapter mirrors that described by Heard (2011) in her analysis of Australian census data. Heard finds that, below the age of 30, cohabitation is most common among those with vocational qualifications. Those on the lower and upper ends of the educational spectrum have lower rates of cohabitation. This research finds that, for first live-in relationship, a similar pattern exists. Comparing direct marriage and cohabitation, those on the upper and lower ends of the educational spectrum are more likely to directly marry, while those in the middle are more likely to choose to cohabit.

This chapter also measured whether the importance of religion (as self-reported) was useful in determining relationship choices. As Australia becomes increasingly secular, this research finds that there is a direct positive relationship between the importance of religion in a person’s life and their decision to cohabit or directly marry. Direct marriage is much more likely when a person indicates that religion is important or very important, while cohabitation is much more likely when religion is considered not important.

This research finds that family background plays an important role in shaping choices around first relationship type. Direct marriage is most likely to occur when young people are from families where one or both parents are born in a non-English speaking country, when the parents have never divorced and where the father has a post-secondary qualification. The impact of family is interesting in this context as it is the setting for a lifetime of relationship modelling. Most young people in this study would have lived with married (not cohabiting) parents, and have not necessarily observed the differences or similarities between couples who cohabit or couples who marry. They have however experienced a high level of parental divorce and this may be crucial in shaping their own attitudes to marriage.

Direct marriage has become increasingly rare as more couples choose to try out their relationship first by cohabiting for a period before marriage. Many experience more than one cohabitation before selecting a partner for marriage. Through this process direct marriage has also become more selective with cohabitation becoming normative and direct marriage occurring at the extremes.

What is the future of LAT relationships and cohabitation among young adults in Australia? LAT relationships for those under 21 will undoubtedly continue along the same lines for many years to come. This is the age at which dating relationships are common. The impact of education and economic factors, such as employment and housing costs, may lengthen the period of ‘dating’ before the first live-in relationship. However, there are probably limits to this prolongation, and these may have already been reached. There is evidence in 2011 and 2012 of a slight drop in the proportion of marriages being preceded by cohabitation (ABS 2013). There is also a very slight increase in the rate of marriage among people aged in their 20s; however, the median age at marriage remains constant. Continued monitoring of these trends is necessary in order to pinpoint the potential limits to the spread of cohabitation and the delay in co-residential relationship formation in Australia.