Living-apart-together (LAT)

relationships in Australia
Anna Reimondos, Ann Evans and Edith Gray

The past few decades have seen substantial However, a growing body of research is now
changes in relationship formation and accumulating on another form of partnership
dissolution patterns in Australia, as in other that is not easily accommodated within this
Western countries, including the postponement tripartite relationship model—that of people
and decline of marriage and the increasing who are in “living-apart-together” (LAT)
popularity of cohabitation. These trends have relationships; that is, those who identify
also led to a change in what demographers and themselves as being in a relationship with
social researchers define as being in a union someone with whom they do not live (Trost,
or relationship. In the past, the distinction 1998). Such relationships are also sometimes
was between those who were married versus described as non-cohabiting or non-residential
those who were single (never married, relationships. Individuals in these unions
separated, divorced or widowed). Today, are essentially “hidden populations”, not
a tripartite model is typically used instead, registered in any official statistics (Borell &
differentiating between those who are single, Ghazanfareeon Karlsson, 2003). This makes
cohabiting or married (Hakim, 2004; Rindfuss it difficult to estimate how common they are,
& VandenHeuvel, 1990; Roseneil, 2006). In but survey evidence from a range of countries
this model, anyone not living in the same suggests that a substantial percentage of the
residence as a partner is classified as being population that would typically be classified
single. This perspective is reinforced by social as single is in fact in a LAT (non-cohabiting
surveys oriented toward collecting data about or non-residential) relationship. In Australia,
households and the relationship of individuals nationally representative surveys indicate that
within those households (Asendorpf, 2008; between 7% and 9% of the adult population
Strohm, Seltzer, Cochran, & Mays, 2010). has a partner who does not live with them.1

Family Matters 2011 No. 87 | 43

 of LAT relationships, it is difficult to know
with certainty if there has been a real rise in
the prevalence of non-residential unions, or
whether this form of relationship has simply
begun to attract more attention than before
(Ermisch & Siedler, 2008). Evidence from a
Swedish poll conducted in 1993, 1998 and 2001
appears to show an increase in the prevalence
of non-residential relationships (Levin, 2004),
and increases between 1982 and 1997 have
also been reported for Japan (Iwasawa, 2004).
However, no increasing trend was found
between 1991 and 2005 in a study using the
German Socio-Economic Panel (Ermisch &
Siedler, 2008). Despite the somewhat mixed
evidence, there are good reasons to believe
that non-cohabiting unions have become
increasingly common in recent years, while at
the same time the recent academic and media
interest in these relationships has also made
Interest in LAT relationships has only recently them more visible.
emerged and there remain questions as to One reason why LATs may be more
how these relationships should be defined prevalent today is the simple fact that a
and accommodated, at both a conceptual and growing proportion of the population is now
theoretical level. As Haskey & Lewis (2006, unpartnered at any point in time (Weston &
p. 38) noted, LAT raises many of the same Qu, 2006). For example, between 1986 and
kinds of questions that were raised when 2001, the percentage of women aged 30–34
cohabitation first came to be widely recognised who were unpartnered (not cohabiting, not
as a distinct form of partnership. These married) increased from 23% to 34% (de Vaus,
questions relate to both the characteristics of 2004), while for men in their early 30s, the
the individuals involved as well as the meaning equivalent rise was from 29% to 41%. As a
of the relationships themselves; whether they larger proportion of the population is not in
are a transitional stage before cohabitation a cohabiting or marital union, the proportion
or marriage, or a completely new form of of the population that is “at risk” of forming a
partnership. Evidence suggests that individuals non-residential relationship has also increased.
enter into non-residential relationships for a Factors contributing to the increased prevalence
range of different reasons throughout the life of individuals who are unpartnered, include
course. socio-economic changes such as the prolonging
To date very little is known about non- of time in education, demographic trends such
cohabiting relationships in Australia, because as increased life expectancy, and increased rates
of the lack of nationally representative survey of relationship dissolution through divorce or
data on this topic. For the first time in Australia, the breakdown of cohabitations (Castro-Martín,
questions on LAT relationships were asked in Domínguez-Folgueras, & Martín-García, 2008;
the 5th wave (2005) of the Household, Income Milan & Peters, 2003; Weston & Qu, 2006).
and Labour Dynamics in Australia (HILDA) At the same time, ideational changes have
survey. We used these data to examine LAT made alternative forms of partnerships more
relationships, with the aim of providing an acceptable in society, and couples who find
estimate of their prevalence, investigating themselves in a relationship with a new partner
the demographic and socio-economic who lives elsewhere may not feel as great a
characteristics of individuals in these unions, social pressure as they would have in the past
and examining how the meaning attached to to settle down together in a common residence
these unions varies across the life course. (Levin, 2004).
The greater availability of quantitative data
Background from social surveys as well as qualitative
studies has also made these relationships more
According to Levin (2004), the LAT relationship visible (Haskey & Lewis, 2006; Trost, 1998).
is a “new family form” that existed in the Until recently, the majority of research on LAT
past but in recent years has become much relationships originated from Scandinavian
more prevalent and visible in society. Due to countries, such as Sweden and Norway (Levin
the lack of historical data on the prevalence & Trost, 1999) where this type of union has

44 | Australian Institute of Family Studies

long been socially recognised and accepted Haskey (2005) used a number of alternative
as a distinct type of relationship (Borrel & ways to try and estimate the “true” number of
Ghazanfareeon Karlsson, 2003). However, LATs from the Omnibus Survey in Britain, such One reason why
with the dramatic increase in interest in LATs as by excluding the relationships of young LATs may be
during the past few years, research is now also adults who were still living at home. more prevalent
accumulating from a growing list of countries,
including France (Beaujouan, Regnier-Loilier,
Since quantitative data on non-residential today is the
unions come from surveys, question wording simple fact that
& Villeneueve-Gokalp, 2009), Germany
plays a very important role in determining
(Asendorpf, 2008), Spain (Castro-Martin et al.,
how such unions are defined and enumerated. a growing
2008), United Kingdom (Duncan & Philips,
2010; Ermisch & Siedler, 2008; Haskey 2005;
As Haskey (2005) recognised, it would be proportion of
impossible to simply ask survey respondents the population is
Haskey & Lewis, 2006), Canada (Milan &
whether they were in a “living-apart-together”
Peters, 2003), the United States (Strohm et al.,
relationship as most people would not now unpartnered
2010) and Japan (Iwasawa, 2004).
understand the term without some explanation at any point
In terms of prevalence, evidence suggests and elaboration. Instead, surveys have used in time.
that a substantial minority of the total adult a variety and combination of terminologies
population is involved in a LAT relationship. to distinguish dating relationships from more
For example, in the US, data from the General committed unions (Strohm et al., 2010).
Social Survey of 1996 and 1998 indicate that Examples include:
6% of women and 7% of men are in a LAT ■■ “Do you have a main romantic involve­
relationship (Strohm et al., 2010), and the ment—a (man/woman) you think of as
2001 Canadian Social Survey produces similar a steady, lover, a partner, or whatever?”
estimates of 8% of the population aged 20 and If yes, respondents are asked if they live
over being in a LAT relationship (Milan & Peters, with their partner. (1996 & 1998 US General
2003). Direct international comparisons of the Social Survey)
prevalence of LAT relationships are difficult to ■■ “Are you in an intimate relationship
make, however, because of the differences in with someone who lives in a separate
the age ranges of the analytical samples, the household?” (2001 Canadian Survey)
dates of the surveys, and the definitions of LAT
used. Due to the relatively recent emergence ■■ “Do you have a steady relationship with a
of scholarly interest in non-cohabiting unions, male or female friend whom you think of
as your ‘partner’, even though you are not
there is still a lack of consensus regarding their
living together?” (1998, 2003 & 2008 British
precise definition.
Household Panel Survey)
■■ “Are you currently in an intimate ongoing
Definition relationship with someone you are not
With regard to definitions, one of the most living with?” (Generations and Gender
important questions in the recent literature is Survey (GGS); 2005 & 2008 HILDA)
where, if anywhere, the boundary between The degree to which these questions are able to
casual dating relationships and more committed exclude less committed relationship varies. The
LAT relationships should be drawn. In general,
there is some agreement that more casual and
fleeting relationships should be differentiated
from more committed non-residential unions,
and often different terms are used to make a
theoretical distinction between the two. For
example, Haskey (2005) termed the former as
“those who have a partner who usually lives
elsewhere”, and the latter as LAT relationships.
Similarly, Trost (1998) used the term “steady
going couples” to identify casual relationships,
as distinct from the more committed LAT
couples. However, in practice, trying to
categorise respondents into one group or the
other is very difficult. Various factors have
been used by researchers as proxy indicators
of the level of seriousness of the relationship.
In their study of young people in Spain, Castro-
Martin et al. (2008) excluded LAT relationships
that had lasted for less than two years, while

Family Matters 2011 No. 87 | 45

 cohabiting in the future is there if and when
circumstances change.
Alternatively, LAT relationships can be more
permanent arrangements that allow for intimacy
but also autonomy and independence, and this
appears to be particularly the case for older
individuals (Levin, 2004). Other reasons for
actively wanting to live apart include the feeling
of not being ready to live with someone, and
having concerns about children (Beaujouan
et al., 2009). Qualitative evidence suggests that
those who are voluntarily living apart include
individuals who have already gone through
a divorce or a relationship breakdown—
experiences that have left them particularly
“risk averse” (de Jong Gierveld, 2004; Levin,
2004; Roseneil, 2006).
While useful, the distinction between LAT
relationships that are involuntary and those
that are more voluntary and permanent
arrangements, might not always be clear cut.
Based on a qualitative study of LATs in the
United Kingdom, Roseneil (2006) suggested that
apart from these two main groups of LATS—the
“regretfully apart” and the ‘gladly apart’—there
last question, which was included in HILDA, is is also a large group of individuals who are
probably one of the strongest as it includes the “undecidedly apart”. People in this group have
terms “intimate” and “ongoing”. Nevertheless, not made a definite choice to cohabit or not.
unlike questions on more objective concepts, Some speak of not being ready or of fearing
such as legal marital status, it is inevitable that cohabitation may ruin their relationship;
that even the clearest question about non- reasons that have also been mentioned in other
residential unions will involve an element of qualitative studies (Haskey & Lewis, 2006). As
subjectivity (Haskey, 2005). Haskey & Lewis noted, fear and perceptions
of risk are important considerations, because
An important theoretical question regarding in many ways the “leap of faith” needed for
LATs relates to the meaning of these moving from a LAT relationship to a cohabiting
partnerships and whether they are a transitory one is greater than the one needed to transition
step taken before entering a live-in relationship, from cohabitation to marriage.
or whether they represent a more permanent
arrangement. A closely related distinction is The purpose of this study was to investigate
whether partners are living apart voluntarily, the nature and pattern of LAT relationships in
through an active choice, or involuntarily due Australia. The first objective was to determine
to constraining circumstances (Levin, 2004). the prevalence of non-residential relationships
Previous research suggests that the meaning in Australia and to describe the key features
of these relationships, including their duration,
The distinction of LAT relationships and the reasons why
the frequency of contact and the geographic
between LAT individuals enter into them, depends very much
distance between partners. We then identified
on what stage of the life course an individual is
relationships that at (Beaujouan et al., 2009; Strohm et al., 2010).
a typology of four types of individuals at
different stages of their life course who were
are involuntary LAT relationships appear to be more in LAT relationships, and investigated how
and those provisional and involuntary among younger the meaning and purpose of non-residential
that are more cohorts. The geographic location of places relationships varied across these four types
of work or study, as well as financial and of people. If the nature and pattern of LAT
voluntary and
housing factors may prevent young people relationships in Australia is similar to that found
permanent from moving into a joint residence with their from research in other Western countries, we
arrangements, partner. Involuntary relationships may also would expect that for younger individuals, a
might not always be the result of having caring responsibilities LAT relationship is likely to be a transitional
for children or elderly parents (Levin, 2004). relationship, or a step towards cohabitation,
be clear cut. While these circumstances prevent moving in while for older individuals it may be a more
together, for these individuals the possibility of voluntary and permanent arrangement.

46 | Australian Institute of Family Studies

Data and method characteristics of LAT respondents compared
to those who are single, cohabiting or
LAT relationships
Data married, using weighted percentages and
summary statistics. Three important aspects can be more
To investigate the prevalence and charac­
teristics of non-residential unions in Australia,
of LAT relationships are also investigated permanent
and compared by age: relationship duration,
we used data from the 5th wave of the
frequency of contact and geographical distance
arrangements
Household, Income and Labour Dynamics in
between partners. In the second section, we that allow for
Australia (HILDA) survey. HILDA is a large-
scale nationally representative longitudinal
use multiple correspondence analysis (MCA) intimacy but also
and cluster analysis to identify different profiles
survey that is conducted on an annual basis autonomy and
of respondents with similar demographic
and interviews all members of a household
characteristics. These different groups of independence,
aged 15 and over. In the 5th wave, several key
questions were included for the first time as part
individuals are then compared in terms of and this appears
differences in their intentions for the future of to be particularly
of Australia’s participation in the international
their relationships.
Generations and Gender Survey (GGS), a cross- the case for older
national longitudinal survey coordinated by the
Population Activities Unit of the United Nations Creating a typology of LATs individuals.
Economic Commission for Europe. To create a typology of individuals in LAT
The question in HILDA asked respondents relationships, we followed the strategy used
who were not married and not living with a by Beaujouan et al. (2009), who analysed
partner, whether they were in an intimate the French GGS data and used multiple
ongoing relationship with someone with correspondence analysis. MCA is a method
whom they were not living. Respondents who of identifying patterns among three or more
answered yes were then asked a series of more categorical variables (Greenacre, 2007),
detailed questions regarding their relationships, which works by converting contingency
including: the month and year the relationship tables into a low-dimensional (typically two-
started, whether a definite decision to live dimensional) space or map.4 The proximities,
apart had been made (and if yes by whom), shown graphically on the map, can then be
the geographic distance to the partner, and the interpreted. In this case, we were interested
frequency of contact. Respondents were also in the interrelationship between four key
asked if they intended to start living with their demographic variables of individuals in
current partner during the next three years, LAT relationships: their sex, their age, the
and if they planned to marry in the future. presence of children, and previous relationship
history. Age is divided into four categories.
It is important to note that the questions on The presence of children is described by a
non-residential partnerships were restricted three-category variable indicating whether
to those who were not married, unlike in the respondent had no children, at least one
the standard GGS questionnaire, where the resident child, or only non-resident children.
possibility that a respondent is married and Finally, previous relationship history was
in a relationship with their spouse but not divided into three categories, consisting of
living with them is included.2 The questions those previously married, previously cohabited
were asked of both heterosexual and same- but never married, and never cohabited or
sex couples, and we have included both types married.
of couples in this study.3 Also, we made no
specific distinction at the outset between more Based on the results of the MCA, Ward’s cluster
and less casual relationships; all living-apart- analysis of the coordinates of the observations
together relationships were considered, even if was used to group together respondents with
they had only been ongoing for a short time. similar characteristics. We identified four
The total sample size in Wave 5 was 12,759 types of individuals with similar demographic
individuals, but as relationship questions were characteristics and, using these groups, we
only asked of respondents aged 18 or over, or investigated how the groups differed in their
less than 18 but not living with parents, we answers to three key questions: whether or not
excluded all those aged less than 18. This they have made a definite decision to live apart,
resulted in a total analytical sample of 12,066 whether they intend to live together within the
respondents, of which 974 were in a LAT next 3 years, and whether or not they intend
relationship. to marry. This allowed us to see whether their
relationship was voluntary or involuntary,
and the degree to which it was seen as a
Method transitional or permanent arrangement. We
The analysis is undertaken in two parts. The used MCA and cluster analysis rather than
first section describes the prevalence and regression methods to examine demographic

Family Matters 2011 No. 87 | 47

 differences in relationship intentions because People with a non-cohabiting partner had a
of the high multicollinearity between many of similar employment status to those who were
the variables in which we were interested.5 cohabiting, and the education profile of the
LATs was also similar to both cohabiting and
married individuals. However, the percentage
Results of LATs who were only educated up to Year 11
Prevalence and characteristics of LATs or below, was lower than for the relationship
status of the other groups. This could be due
The distribution of the HILDA sample according to a cohort effect, since younger people have
to their relationship status (single, married, higher education levels than older cohorts;
cohabiting or LAT) by age, is shown in Figure 1. however, it might also be that LAT relationships
The graph shows a clear pattern between age are simply more prevalent among those with
and relationships status, with the proportion higher educational levels, as other studies from
that were in a LAT or cohabitating relationship the United Kingdom, Spain and Germany have
declining with age, and the proportion in a found (Castro-Martin et al., 2008; Ermisch &
marriage increasing.6 Of those aged 18 and Siedler, 2008; Haskey & Lewis, 2006). A possible
over, 36% were not married or cohabiting, a reason for this is that individuals with higher
figure that is similar to the 39% estimated by education and occupational statuses are more
the 2006 Census (Australian Bureau of Statistics likely to have jobs that require a degree of
[ABS], 2009). The overall percentage of the travel and mobility, and at the same time they
sample that was in a non-cohabiting union was are more able to afford two separate residences
9%, which represents 24% of those who were (Haskey & Lewis, 2006). Castro-Martin et al.
not cohabiting or married. (2008) also suggested that among younger
individuals in Spain LAT arrangements may suit
Apart from age, relationship status was also those who prioritise a professional career.
related to several other key demographic and
socio-economic variables, as shown in Table 1. Duration, frequency of contact and
Compared to those who were single, cohabiting
or married, those in a LAT relationship differed distance between partners
on several key characteristics. They were less At the time of the survey, the majority of LAT
likely to have children compared to those who relationships had, on average, been ongoing
were married, cohabiting or single, and they for only a short amount of time;7 the mean
were also less likely to have ever been married. duration of these relationships was 2.4 years,
Their marital and fertility history is, of course, and the median was 1.5 years. These averages,
closely related to their young age profile. however, hide quite substantial variations in
relationship durations. While the majority of
respondents (40%) had began their relationship
LAT Cohabiting Married Single less than 12 months before the survey, a
100
substantial 28% were in a relationship that had
lasted for 3 years or more. The cross-sectional
90 nature of the data makes it difficult to draw
any conclusions about the timing of transitions
80
out of non-cohabiting relationships, but the
Percentage of respondents

70 results seem to indicate that after one or two

years, individuals in non-resident relationships
60
commonly experience some transition, either
50 by ending the relationship or by starting to live
together. How long a LAT relationship had been
40
ongoing was closely related to a person’s age,
30 as shown in Figure 2. In particular, those aged
45 and over stood out, as nearly half of the
20
people in this age group were in a relationship
10 that had lasted 3 years or more, compared to
less than a third of respondents in all the other
0
18 22 26 30 34 38 42 46 50 54 58 62 66 70 74 younger age groups.
Age Another important characteristic of LAT
relationships that we considered was how
Source: HILDA Wave 5 frequently the partners met. The frequency
Figure 1 Relationship status, by age (18–75+ years) of contact between partners gives us some
indication of the importance of a LAT partner to

48 | Australian Institute of Family Studies

 Table 1 Demographic characteristics of sample, by relationship status
Single LAT Cohabiting Married Total
% (weighted)
Age group (years)
18–24 25 44 16 1 13
25–29 10 14 19 5 9
30–34 8 14 19 10 10
35–39 7 6 11 11 10
40–44 7 6 11 12 10
45+ 44 16 25 61 49
Number of children
0 55 73 50 11 32
1+ 45 27 50 89 68
Ever married
Yes 43 21 28 100 71
No 57 79 72 0 29
Employment
Employed 55 82 79 62 64
Unemployed 5 4 3 1 3
Not in the labour force 39 14 18 36 34
Highest education
University 16 22 23 22 21
Certificate/diploma 26 30 34 32 30
Year 12 21 30 17 12 16
Year 11 or below 37 18 26 34 33
Total percentage 100 100 100 100 100
Total N (unweighted) 3,290 974 1,509 6,293 12,066

Source: HILDA Wave 5

an individual’s daily life. We found that despite

not sharing the same residence, the frequency 0–1 years 1–2 years 2–3 years 3+ years
of contact between partners was very high, 100
with around 75% of individuals meeting their
partners at least three times a week, and many
of these on a daily basis. 80

As with the duration of the relationship, the

frequency of contact also showed some 60
Percentage

variation by age, as shown in Figure 3 (on

page 50). The association with age was 40
negative, with those aged under 35 meeting
their partners most often. At these ages, the
two partners may be attending the same school, 20
university or workplace, which would allow
them to meet their partner on a frequent basis.
0
The relatively less frequent contact among 18–24 25–34 35–44 45+
those over 35 could also be related to the fact Age group
that these older respondents are also the ones
who are more likely to have resident children, Source: HILDA Wave 5
and would therefore experience greater time
Figure 2 Duration of LAT relationship, by age group
constraints. Nevertheless, even among those

Family Matters 2011 No. 87 | 49

 aged 45 and over, around 80% still saw their same state. Only a minority were in a long-
partner on a weekly basis. distance relationship with a partner who lived
in another state (2%) or overseas (5%). The
A key factor regulating the frequency of close physical proximity between partners’
contact between partners was how close they residences was also indicated by the travel
lived to each other. As might be expected from time between residences. Around one quarter
the high frequency of contact, the majority of of individuals lived less than 10 minutes away
people with a non-residential partner lived from their partners, with the median time being
very close by to their partners. Nearly three- 20 minutes.
quarters lived in the same city as their partner, The relationship between age and geographic
and a further 15% lived in different cities in the location between partners is shown in Figure 4.
Out of all age groups, those aged 45 and over
were the most likely to be living in different
cities, which would also explain the lower
3+ times a week 1–2 times a week < 1 week frequency of contact mentioned above. It is
100 interesting to note that among the 25–34 year
age group, just under 10% had a relationship
with someone living abroad. It could be
80 speculated that this is related to the relatively
high degree of travel undertaken at these ages
60 (ABS, 2008). These respondents may have met
Percentage

a partner while travelling or working overseas,

or their partner may have moved overseas for
40 travel or work reasons.

20 Typology of LATs
Our next step was to identify a typology of
0 individuals in LAT relationships. From the MCA
18–24 25–34 35–44 45+ and cluster analysis, we identified four main
Age group types of individuals, whose characteristics are
shown in Table 2. We called the first group
Source: HILDA Wave 5 the “Under-25s” because they were primarily
made up of individuals aged 18–24. This group
Figure 3 Frequency of contact between partners, by age group was relatively homogenous, having had no
children, no previous history of marriage, and,
for the most part, no history of cohabitation.
The second group we termed “Young adults,
previously de facto”. As the name implies, this
group was primarily made up of young adults
Same city/town Same state Different state Overseas
between the ages of 25 and 34, the majority
100 of whom were childless and had no marriage
history, but usually had experienced at least
80 one cohabitation in the past. The third type was
the “Single parents”, who were older people,
typically over 30, most of whom had been
60 married and had at least one child resident
Percentage

in the household. Finally, we identified the

40 “Older, previously married” group, which was
also relatively homogeneous, consisting mainly
of those aged 45 and over who had previously
20 been married. Our typology of individuals was
very similar to the one derived from the French
0 GGS (Beaujouan et al., 2009).
18–24 25–34 35–44 45+
Age group Intentions to cohabit or marry
Based on our typology of individuals, we
Source: HILDA Wave 5
then examined group differences in whether
Figure 4 Geographic distance between partners, by age group a definite decision had been made to live
apart, and in their future intentions to cohabit

50 | Australian Institute of Family Studies

 Table 2 Demographic characteristics of respondents in LAT relationships, by type of individual
Young
Older,
adults, Single
Under-25s previously Total
previously parents
married
de facto
%
Sex
Male 63 28 48 65 53
Female 37 72 52 35 47
Age group
18–24 75 15 2 0 44
25–29 13 32 5 0 14
30–34 9 35 12 0 14
35–39 2 10 15 0 6
40–44 1 9 20 0 6
45+ 0 0 46 100 16
Number of children
0 100 74 19 0 73
1+ 0 26 81 100 27
Ever married
Yes 0 0 73 100 79
No 100 100 27 0 21
Employment
Employed 85 78 82 69 82
Unemployed 5 4 2 2 4
Not in the labour force 10 17 17 29 14
Highest education
University degree 20 28 23 19 22
Certificate/Diploma 25 39 33 38 30
Year 12 40 21 17 12 30
Year 11 or below 15 12 28 31 18
Total percentage 100 100 100 100 100
Total N (unweighted) 474 184 219 97 974
Within-class variance 0.06 0.25 0.42 0.16 Despite not
Source: HILDA Wave 5
sharing the same
residence, the
or marry. We found clear differences between because of constraints such as lack of financial frequency of
the group responses, including whether they resources that prevented them from moving in contact between
had made a definite decision to live apart, as with their partner. partners was
shown in Table 3. For example, while over 70%
of the older respondents who had previously When there had been a definite decision to live very high, with
been married had made a positive decision to apart, most people indicated that it was a joint around 75%
live apart, this was the case for less than half of decision between them and their partner. The of individuals
the under-25s. single parents were most likely to state that the
decision to live apart was solely theirs, followed meeting their
To some extent, a definite decision to live apart
could be taken as implying that the arrangement by the older group. While the responses of the partners at least
of not living in the same household was one single parents are not surprising, it is interesting three times a
of choice rather than constraint. However, this that in the other groups where the decision was week, and many
may not always be clear-cut, because young not joint, individuals usually stated that it was
their decision alone, even though we would
of these on a
adults still living at home may have stated that
they had made a definite decision to live apart, expect the decision to be roughly equally daily basis.
as opposed to a preferred choice, but only divided between the two partners.

Family Matters 2011 No. 87 | 51

 Table 3 Intentions for the future of relationship, by type of individual
Young adults, Older,
Single
Under 25s previously de previously Total
parents
facto married
Definite decision to live apart
Yes 48 61 67 73 57
No 52 39 33 27 43
Total (%) 100 100 100 100 100
Total (N) 474 184 219 96 973
Whose decision to live apart
Respondent 11 15 23 17 16
Respondent’s partner 2 4 5 7 4
Both respondent and
87 81 72 76 80
partner
Total (%) 100 100 100 100 100
Total (N) 229 111 147 70 557
Intend to live together within next 3 years
Yes 69 79 53 32 64
No 31 21 47 68 36
Total (%) 100 100 100 100 100
Total (N) 468 178 210 90 946
Likelihood of marrying/re-marrying
Unlikely/very unlikely 6 12 44 68 22
Not sure 23 31 24 17 24
Likely/very likely 71 57 32 16 54
Total (%) 100 100 100 100 100
Total (N) 473 182 216 97 968

Source: HILDA Wave 5

Turning to people’s intentions regarding the It is also interesting to note that there was not
future of their relationships, nearly two-thirds always a close link between having made a
of respondents planned to live together within definite decision to live apart and intentions
the next 3 years, although there was large to not live together at all. For example, while
degree of variation in responses (Table 3). The 61% of young adults (previously de facto) had
young adults were the group with the highest made a decision to live apart, around 79% did
stated intentions of living with their partner, at intend to move in together within the next 3
79%, while the lowest intentions were found years.
among the older group, at 32%.
Respondents were also asked about their
In general, it is difficult to tell whether those plans for marriage in the future. There was no
who stated that they did not intend to live explicit mention in this question whether the
together in the next 3 years were just uncertain future marriage was to the current LAT partner
There was not if their relationship would continue that far or to a hypothetical future partner; we assumed
into the future, or whether their answer was that the majority would answer with respect
always a close more an indication of a preference to maintain to their current partner. As with the intention
link between the status quo and stay in the relationship in to cohabit, responses to the marriage question
having made a the long term but continue living in separate also varied greatly among the groups. Among
residences. Given the earlier results regarding the under-25s group, just over 70% thought
definite decision the age pattern of duration, it may be speculated that they were likely or very likely to marry in
to live apart and that for young adults a negative intention the future, and attitudes towards marriage were
intentions to reflects an uncertainty about the future of the also positive among young adults who had
relationship, while for older adults—who had previously been de facto. On the other hand,
not live together
relationships of the longest duration—it could single parents and older respondents had much
at all. indicate a preference to keep the current living lower intentions of marrying, a result that was
arrangements. also found by Ermisch & Siedler (2008). Over

52 | Australian Institute of Family Studies

two-thirds of the older respondents said they residence. Around half of the single parents
were unlikely or very unlikely to remarry in did, however, envisage living with their partner Older
the future. in the next 3 years. At this time, the resident
children may have grown accustomed to the respondents,
partner, or they may have grown up and left most of whom
Discussion the household. were widowed
The results from the analyses of HILDA
data closely resemble the ones from other
A high percentage of young adults who or divorced, were
had previously cohabited intended to start
international studies. In particular, we find cohabiting with their partner in the next 3
the most likely to
that older respondents, most of whom were years, and also to marry in the future. This be “voluntarily”
widowed or divorced, were the most likely to be group may feel the greatest normative pressure living-apart-
“voluntarily” living-apart-together and to have to consolidate their relationship by living in
little intention to transition into cohabitation. together and
a common residence. For those under 25,
While we do not know the reasons behind the single parents, and the older, previously to have little
the choice, the wish to maintain a degree of married, couples, the pressure to move in with intention to
independence and autonomy is likely to be their partner is unlikely to be felt as strongly.
an important consideration (Beaujouan et al., transition into
Indeed, these groups may even have felt a
2009). Qualitative research of LAT relationships social pressure not to live with their partner. cohabitation.
in later life in other countries highlights that
for the elderly, important concerns appear The under-25s groups was more evenly
to centre around the practicalities of sharing divided in terms of whether a definite decision
living quarters with someone else and having had been made to live apart. In this group, we
to adjust to another person’s habits, and the may be picking up a substantial proportion of
wish to remain autonomous and maintain casual and fleeting relationships. For the more
or continue relationships with children and committed partners, the arrangement may be
grandchildren (de Jong Gierveld, 2002). more a matter of circumstances and practical or
financial constraints rather than choice. At this
The single parents most closely resembled age, and with no previous experience of living
the older respondents in their decision to live with a partner, they may also not feel ready to
apart and their future plans for cohabitation. take the step to move in with their partner.
Again, we do not know the reasons behind the
decision, though it is possible that they did not
want to disrupt the home environment of their
Conclusion
resident child(ren) by bringing a new partner Changing demographic trends mean that a
into the home or by moving into another substantial proportion of the population is

Family Matters 2011 No. 87 | 53

 it is also important to learn more about their
intimate relationships, because non-residential
partners may be an important source of
instrumental and emotional support, especially
for the elderly who are living alone.
It is important to not only understand more
about these relationships in their own right,
but also a greater understanding of why new
relationship types such as LATs are formed can
provide some insight into reasons for changing
relationship trends, such as the postponement
and avoidance of marriage (Casper, Brandon,
DiPrete, Sanders, & Smock, 2008; Strohm et al.,
2010). However, to understand the nature of
live-apart-together relationships at all stages
of the life course, more quantitative and
qualitative research is needed. At the moment,
our understanding of these relationships is
now not living with a partner. According
limited by the cross-sectional nature of most
to the 2006 Census, in Australia 4.6 million
quantitative studies. Longitudinal data on
people aged 20 and over, or nearly a third
LAT relationships would allow us to study
of the adult population, were not living with
their duration, and their possible eventual
a partner or spouse and could therefore be
development into separations or cohabitations.
classified as being unpartnered (ABS, 2007).
We estimated from the HILDA data that around
24% of the single population was in fact in a Endnotes
relationship, albeit not living with their partner.
1 In HILDA Wave 5 (2005), the weighted percentage
This translates to over 1.1 million Australians in of those aged 18 and over and with a non-resident
living-apart-together relationships. We suggest partner was 9%. In Wave 4 (2006) of the Negotiating
that it is important to understand more about the Life Course survey, the weighted percentage of
these partnerships, as the lives of people who those aged 18–65 and not living with their partner
are truly single, compared with people who was 7%. In the Family, Social Capital and Citizenship
survey, 7% of the sample had a non-resident partner
have a non-resident partner, are likely to be
(unweighted).
different in many respects. Several authors
2 In the literature, there is no standard treatment
have also predicted that LAT relationships are of married couples in living-apart-together
going to become more common in the future. relationships. Sometimes married couples are
Reasons for this include the ones discussed included in the definition of LAT unions (Levin &
earlier, such as the continuation of demographic Trost, 1999), and other times excluded (Haskey,
trends of increased life expectancy, increased 2005). There is general consensus, though, that LAT
relationships do not include so-called “commuter”
rates of marital dissolution and the rise of
marriages/cohabitations, where the couple
cohabitations. Also important may be increased maintains one household, but one partner lives
gender equality and the rise of dual-career elsewhere for periods of time due to work reasons
couples, and cases where working women (Levin & Trost, 1999).
For younger are less able to relocate for their partner’s job 3 There were 18 same-sex couples in the HILDA
individuals, it (Levin, 2004; Castro-Martin et al., 2008). sample (< 2% of those were in LAT relationships).

is important For younger individuals, who are known

4 We used the XLSTAT software to perform the
multiple correspondence and cluster analysis
to identify the to be moving out of the parental home at <www.xlstat.com>.
increasingly later ages (de Vaus, 2004), it is
constraints they 5 An alternative analysis strategy to that outlined
important to identify the constraints they face would be to have each of the three key topics of
face in setting in setting up a common residence with their interests (whether relationship is due to a definite
up a common partner, in particular in relation to financial decision to live apart, intention to live together
in next 3 years, and intention to marry) as the
residence with and housing factors. More qualitative research
dependent variable and to see how key demographic
would also be of interest in order to understand
their partner, young people’s attitudes towards establishing
and socio-economic characteristics—such as age,
number of children and employment—influenced
in particular a common residence with a partner, and the the dependent variable. This strategy was not used
because of the high degree of multicollinearity
in relation to degree to which non-residential relationships
between age and the other independent variables;
among young people are related to
financial and individualistic values, risk aversion or fear of
for example, younger respondents are much
more likely to have never had a previous live-in
housing factors. commitment, as has been suggested (Heath relationship and to not have children, while the
& Cleaver, 2003). Among older individuals, opposite is true for older respondents. This made it

54 | Australian Institute of Family Studies

 difficult to separate out the effects of age from the Ermisch, J., & Siedler, T. (2008). Living apart together. In
other variables of interest. M. Brynin, & J. Ermisch (Eds.), Changing relationships
6 Since these are cross-sectional results, the pattern is (pp. 29–43). New York: Routledge. A greater
influenced by both cohort and age effects. Greenacre, M. (2007). Correspondence analysis in understanding
7 Fifty-nine respondents did not know the month the practice. London: Chapman & Hall/CRC. of why new
relationship started, but knew the year. In these Hakim, A. (2004, 15–17 September). Social marital
cases, the month was imputed to June. In addition, status in Australia: Evidence from 2001 Census. relationship
two had missing information on the duration of Paper presented at the 12th Biennial Conference types such as
the relationship, as they did not know the year the of the Australian Population Association, Canberra.
relationship started. Retrieved from <www.apa.org.au/upload/2004–7D_ LATs are formed
Hakim.pdf>. can provide
References Haskey, J. (2005). Living arrangements in contemporary
Britain: Having a partner who usually lives elsewhere
some insight
Asendorpf, J. B. (2008). Living apart together: Alters-
and living apart together (LAT). Population Trends, into reasons
und Kohortenabhängigkeit einer heterogenen
35–45.
for changing
Lebensform. Kölner Zeitschrift für Soziologie und Haskey, J., & Lewis, J. (2006). Living-apart-together in
Sozialpsychologie, 60(4),749–764. Britain: Context and meaning. International Journal relationship
Australian Bureau of Statistics. (2007). 2006 Census
of Law in Context, 2(1), 37–48. trends,
tables: Census of Population and Housing: Social Heath, S., & Cleaver, E. (2003). Young, free and
single? Twenty-somethings and household change.
such as the
marital status by age by sex for time series (Cat. No.
2068.0). Canberra: ABS. Retrieved from <www.abs. Houndmills, Basingstoke, Hampshire: Palgrave postponement
gov.au/websitedbs/d3310114.nsf/Home/census>. Macmillan.
and avoidance of
Australian Bureau of Statistics. (2008). Migration, Iwasawa, M. (2004). Partnership transition in
Australia 2006–07 (Cat. No. 3412.0). Canberra: ABS. contemporary Japan: Prevalence of non-cohabiting marriage.
Retrieved from <www.abs.gov.au/AUSSTATS/abs@. couples. Japanese Journal of Population, 2(1), 76–92.
nsf/DetailsPage/3412.02006-07?OpenDocument>. Levin, I. (2004). Living apart together: A new family
Australian Bureau of Statistics. (2009). Couples in form. Current Sociology, 52(2), 223–240.
Australia (Cat. No. 4102.0). Canberra: ABS. Retrieved Levin, I., & Trost, J. (1999). Living apart together.
from <www.ausstats.abs.gov.au/ausstats/subscriber. Community, Work & Family, 2(3), 279 –294.
nsf/0/6F761FF864FAA448CA2575830015E923/$Fi
Milan, A., & Peters, A. (2003). Couples living apart.
le/41020_couples.pdf>.
Canadian Social Trends, Summer, 2–6.
Beaujouan, E., Regnier-Loilier, A., & Villeneueve-
Rindfuss, R. R., & VandenHeuvel, A. (1990). Cohabitation:
Gokalp, C. (2009). Neither single, nor in a couple: A
A precursor to marriage or an alternative to being
study of living apart together in France. Demographic
single? Population and Development Review, 16(4),
Research, 21(4), 75–108.
703–726.
Borell, K., & Ghazanfareeon Karlsson, S. (2003).
Roseneil, S. (2006). On not living with a partner:
Reconceptualizing intimacy and ageing: Living apart
Unpicking coupledom and cohabitation. Sociological
together. In S. Arber, K. Davidson, & J. Ginn (Eds.),
Research Online, 11–3. Retrieved from <www.
Gender and ageing: Changing roles and relationships
socresonline.org.uk/11/3/roseneil.html>.
(pp. 47–62). Philadelphia: Open University Press.
Strohm, C., Selzer, J. A., Cochran, S. D., & Mays, V.
Casper, L. M., Braddon. P. D., DiPrete, T. A., Sanders,
(2010). “Living apart together” relationships in the
S., & Smock, P. J. (2008). Rethinking change and
United States. Demographic Research, 21(7), 177–214.
variation in unions. In P. I. Morgan, C. Bledsoe, S.
M. Bianchi, P. L. Chase-Landsdale, T. A. DiPrete, V. J. Trost, J. (1998). LAT relationships now and in the
Hotz et al., (Eds.), Explaining family change: Final future. In K. Matthijs (Ed.), The family: Contemporary
report (pp. 59–100). Durham: Duke University. perspectives and challenges (pp. 209–222). Louvain,
Retrieved from <www.soc.duke.edu/~efc/>. Belgium: Leuven University Press.
Castro-Martín, T., Domínguez-Folgueras, M., & Martín- Weston, R., & Qu, L. (2006). Family statistics and trends:
García, T. (2008). Not truly partnerless: Non- Trends in couple dissolution. Family Relationships
residential partnerships and retreat from marriage in Quarterly, 2, 9–12.
Spain. Demographic Research, 18(16), 436–468.
Anna Reimondos is a Research Assistant, Dr Ann
de Jong Gierveld, J. (2002). The dilemma of repartnering: Evans is a Fellow, and Dr Edith Gray is a Fellow,
Considerations of older men and women entering all at the Australian Demographic and Social Research
new intimate relationships in later life. Ageing Institute, Australian National University.
International, 27(4), 61–78.
This paper was funded by the Australian Research
de Jong Gierveld, J. (2004). Remarriage, unmarried
Council (DP0772544 & DP110105019) and uses
cohabitation, living apart together: Partner
confidentialised unit record data from the Household,
relationships following bereavement or divorce.
Income and Labour Dynamics in Australia (HILDA)
Journal of Marriage and Family, 66(1), 236–243.
Survey. The HILDA Project was initiated and is funded
de Vaus, D. A. ( 2004). Diversity and change in by the Department of Families, Housing, Community
Australian families. Melbourne: Australian Institute Services and Indigenous Affairs (FaHCSIA) and is
of Family Studies. Retrieved from <www.aifs.gov.au/ managed by the Melbourne Institute of Applied
institute/pubs/diversity/main.html>. Economic and Social Research (MIAESR). The findings
Duncan, S., & Philips, M. (2010). People who live apart and views reported in this paper, however, are those
together (LATs): How different are they? Sociological of the authors and should not be attributed to either
Review, 58(1), 112–134. FaHCSIA or the MIAESR.

Family Matters 2011 No. 87 | 55