Factors associated with housing stability for Aboriginal families in South Australia: a prospective cohort study

Abstract

Stable, affordable and safe housing is a key determinant of health and wellbeing. Evidence from an Australian population-based cohort of 344 Aboriginal mothers and children aged 5-9 years shows that tenure type, financial difficulties and women’s experience of partner violence are key drivers to housing stability. Stable housing was defined as moving house ≤ two times in the previous five years. Women had higher odds of stable housing if they had not reported partner violence in the last year (Adj. OR = 2.2, 95%CI 1.1–4.3) and reported fewer financial difficulties (Adj. OR= 2.2, 95%CI 0.9–5.4). Women in private rental had lower odds of stable housing (Adj. OR = 0.4, 95%CI 0.2–0.7) compared to women in government subsidised rental housing. Understanding factors that support housing stability is crucial to ensure the best outcomes for young Aboriginal children and families at a critical period in the life course.

Keywords:

• Stability
• Aboriginal
• families
• residential mobility
• moving house

Introduction

There is global evidence that low income and marginalised populations move house more frequently, are more likely to live in disadvantaged areas and experience involuntary or forced moves (Baker, Bentley, et al., 2016; Cohen & Wardrip, 2011; Ong et al., 2022; Phinney, 2013). For example: a rental lease could be cut short, house payments become unaffordable or family circumstances could change through job loss, family violence, separation or divorce (Baker et al., 2019; James et al., 2021; Mallett et al., 2011). Housing disadvantage and higher levels of homelessness are experienced by First Nations people in countries that share histories of colonisation and dispossession, including New Zealand Māori, North American Indian, Métis and Inuit in Canada and Australian Aboriginal and Torres Strait Islander peoples (Anderson & Collins, 2014; Cohen & Wardrip, 2011; Snyder & Wilson, 2015). Past government policies causing displacement, loss of land and language, continue to impact First Nations people (UN Habitat, 2009). A recent systematic review (Nikolof et al., 2023) examining housing mobility in First Nations families from Australia, New Zealand and Canada found a significant number of families experience high mobility. Housing barriers faced by families included; insufficient housing availability, overcrowding, poor housing conditions, a lack of affordability, long waiting lists for government subsidised housing and discrimination in the private rental market (Nikolof et al., 2023). The review concluded that there is an urgent need to understand underlying reasons for high mobility and impacts on the wellbeing and development of First Nations children.

A stable home can act as a secure base, enabling children and their families to maintain social connections, and have consistent access to support services and employment (Anderson et al., 2014; Baker et al., 2016; Snyder & Wilson, 2015). Importantly, stable housing gives parents a sense of autonomy, certainty and control that leads to lower levels of stress and better quality of life (Australian Institute of Health and Welfare, 2020; Grant et al., 2017; Li et al., 2022; Taylor & Edwards, 2012).

Children’s development in this paper is understood through an ecological, developmental systems perspective, where the child is an active player within increasingly distal ecological systems of family, community, culture and society (Bronfenbrenner & Morris, 1998).Research suggests that early childhood is a sensitive time where high mobility is more likely to impact children’s development and wellbeing (Giles et al., 2013; Lawrence & Molborn, 2016; Schmitt & Lipscomb, 2016; Taylor & Edwards, 2012; Ziol-Guest & McKenna, 2014). Young children with high mobility can experience multiple changes in contexts such as family and kin relationships, community, and education settings including childcare, kindergarten and school, which in turn can impact on health and development (Jelleyman & Spencer, 2008; Simsek et al., 2021).

The housing circumstances of Aboriginal¹ families in Australia differ greatly from those of non-Aboriginal Australians (Andersen et al., 2018). Aboriginal households are larger in size, more likely to include a lone parent and frequently comprise multi-generational households that include members of extended family and kinship networks (Hewitt & Walter, 2022). There is evidence that cultural factors contribute to temporary mobility, with Aboriginal families more likely to visit other family members to maintain kinship ties or stay connected to Country² (Memmott et al., 2006; Tedmanson et al., 2022).

Stable and affordable housing is out of reach for many Aboriginal families in Australia. Aboriginal households are less likely to own their home, more likely to be overcrowded, and more likely to live in rental accommodation (South Australian Aboriginal Housing Strategy, 2021). Each of these factors contribute to increased levels of housing mobility and homelessness (Andersen et al., 2018). Recent literature has found tenure type, poverty, poor housing quality, overcrowding and a lack of affordable housing are key contributors to housing mobility in Australian Aboriginal communities (Andersen et al., 2017; Australian Government Department of Social Services, 2015; Chando et al., 2020; Lowell et al., 2018).

In Australia, rising house costs, a decrease in public housing (government subsidised short and long-term rental) and a low supply of affordable housing options have all added to pressures experienced by families (Muir et al., 2020; Robinson & Adams, 2008). Public housing (also called social housing) is provided by State and Territory governments or community sector organisations, to people on very low incomes, who have experienced homelessness, family violence or have other complex needs (Australian Housing and Urban Research Institute, 2023; Brown et al., 2023). However there is a decreased supply of public housing and growing numbers of households experiencing housing affordability issues. As a response to high demand, government policies ration public housing through tighter eligibility rules resulting in fewer families living in public housing, larger waiting lists and longer waiting times (Muir et al., 2020). Census data indicates that, between 2001 and 2021, the proportion of Aboriginal households that were living in public housing fell from 32% in 2001 to 27% in 2011, and 18% in 2021. Conversely Aboriginal people living in private rental rose from 28% in 2001 to 30% in 2011 and 35% in 2021 (Australian Institute of Health and Welfare, 2023). These policies have impacted disproportionately on younger Aboriginal people. Despite Aboriginal people representing 3.8% of the total Australian population, Aboriginal people accounted for 28% of all users of Specialist Homelessness Services in 2022 (increasing from 22% in 2012). In 2022, 50% of Aboriginal Specialist Homelessness Services clients were aged under 25 compared to 39% of non-Aboriginal clients (Australian Institute of Health and Welfare, 2023).

In addition, the Aboriginal population has a younger age structure than the non-Aboriginal population, for example, in 2021 one-third (33.1%) of Aboriginal Australians were aged under 15 years compared with 17.9% of non-Aboriginal Australians (Australian Bureau of Statistics, 2021). By 2031, the Aboriginal population is projected to grow by 2.0 per cent per year compared to 1.5 per cent for the total Australian population (Australian Bureau of Statistics, 2019). This growth is expected to place greater strain on housing supply and demand for housing services (South Australian Aboriginal Housing Strategy, 2021). Improved housing is a key target of Closing the Gap, a national strategy to overcome entrenched inequality experienced by Aboriginal communities. The strategy was developed in partnership between Australian governments and the Coalition of Aboriginal and Torres Strait Islander Peak Organisations (Australian Government Productivity Commission, 2023).

Despite key differences in the housing circumstances between Aboriginal Australians and non-Aboriginal Australians, most mobility research reports on data from majority group populations and non-Indigenous population groups (Garboden et al., 2017; Hewitt & Walter, 2022; Simsek et al., 2021). Where comparisons are made for Aboriginal Australians, the evidence tends to take a deficit approach and focus on a limited range of indicators and targets that may not reflect Aboriginal perspectives and priorities (Fogarty et al., 2018; Tedmanson et al., 2022). There is limited Aboriginal community driven research, where local community lead the design of questions, the collection, analysis and interpretation of findings (Biles et al., 2022; Harfield et al., 2020).

Whilst the majority of Australian Aboriginal families now live in urban areas (79%) (ABS, 2016) most of the existing literature focuses on remote communities (Dockery & Colquoun, 2012; Lowell et al., 2018) or temporary inter or intra-community mobility within a specific area (Habibis et al., 2011; Memmott et al., 2006). There is a significant gap in research relevant for the broader Aboriginal population who predominantly live in urban and regional areas rather than remote communities. Given the current housing market in Australia, evident housing disadvantage and the growing number of Aboriginal families with young children, there is an urgent need to understand factors associated with stable and safe housing.

This paper examines the housing experiences of families with children aged 5-9 years living in urban, regional, and remote communities in a prospective population-based cohort. The Aboriginal Families Study was conducted in partnership with Aboriginal and Torres Strait Islander communities in South Australia (Buckskin et al., 2013); and over-seen by a long-standing Aboriginal Governance Group. This paper was conducted by a research team comprised of four Aboriginal and two non-Aboriginal researchers with experience in collaborative Aboriginal and Torres Strait Islander health research.

Aboriginal Families Study data from 344 Aboriginal families in South Australia were used to: 1) examine housing stability; 2) describe the sociodemographic and social characteristics of families living in stable housing; 3) explore the relationship between housing tenure, financial difficulties and women’s experience of partner violence and housing stability.

Methods

The Aboriginal Families Study is a prospective cohort study investigating the health and wellbeing of 344 Aboriginal children and their mothers living in urban, regional, and remote areas of South Australia (Buckskin et al., 2013; Weetra et al., 2019). The study builds on consultations with Aboriginal communities across South Australia (2007-2008). The study was conducted in collaboration with the Aboriginal Health Council of South Australia, the peak body representing Aboriginal community-controlled health organisations in South Australia. An Aboriginal Governance Group was established prior to the commencement of the study and continues to work closely with the research team through all stages of the research, keeping community and policy goals in mind (Buckskin et al., 2013). Two waves of data collection have been completed to date. Women completed a structured questionnaire when the children in the study were 4 to12 months of age (2011 to 2013). Wave 2 follow-up (mid-2018 to 2020) was timed to coincide with the study children commencing primary school. At each wave, a team of Aboriginal researchers invited women to complete a structured questionnaire as an interview or to self-complete the questionnaire if preferred. Women were recruited through a range of services and community networks. These included: the community connections and networks of the Aboriginal research team (41%), referral by an Aboriginal health service (20%), referral by a public hospital (16%), referral to the study by other participants (15%), and referral by other community-based agencies (8%) (Buckskin et al., 2013). Comparisons with South Australian perinatal data for births to Aboriginal women in 2011 showed that participants were largely representative in relation to maternal age, infant birthweight and gestation, but slightly over-represent women having their first child (34.3% vs 4.2% in routinely collected data (Brown et al., 2016).

Participants

Women were eligible to participate in the study if they gave birth to an Aboriginal and/or Torres Strait Islander child in South Australia between July 2011 and June 2013 and were aged 14 years or older at the time the child was born. If the study child was not living with their mother at wave 2, the child’s primary carer (e.g. father, other family member, foster carer) was invited to complete a modified version of the questionnaire. Informed consent was obtained from participants by Aboriginal researchers. Further details regarding community consultation, partnership arrangements, and study procedures are available in other papers (Buckskin et al., 2013).

Measures

Housing stability

Women were asked ‘How many times have you moved in the last 5 years? (Number of times you have moved).’ For the purposes of this analysis women were grouped into two categories; (1) stable housing - those who moved two or less times in the previous five years; (2) unstable housing - those who moved three or more times in the previous five years. The definition of stable housing was based on a meta-analysis of 844 studies published between 1989 to 2020 which reported three or more moves in a child’s lifetime was more than twice as detrimental to child health as low mobility (one to two moves) (Simsek et al., 2021). Previous studies have found children who experienced more than three moves in the first five years (Ziol-Guest & McKenna, 2014) and the first seven years (Gaylord et al., 2018) experienced poorer outcomes compared to children who experienced less than three moves over the same period.

Socio demographic factors

Information on women’s age, relationship status, employment, highest level of education, study child’s age, number of adults and children in the house was collected. Metropolitan, regional and remote location was classified based on postcodes and the Australian Bureau of Statistics Remoteness Structure (Australian Bureau of Statistics, 2016). Information was also collected on a range of contextual factors including the Aboriginal language or community group women come from, if they speak any Aboriginal languages and family members involved in parenting the study child.

Social factors

Women were asked ‘What sort of place are you living in? Is it a:’ ‘house’, ‘unit or flat’, ‘hostel’, ‘hotel or motel’, ‘caravan park’, ‘women’s shelter’ or ‘somewhere else (please say where)’. In addition, women were asked ‘How long have you lived in the place where you live now? (Number of years’ or ‘Number of months)’. To assess house tenure type, they were asked ‘Is the place where you live now: –-’. ‘rented from government housing authority (Housing Trust³)’, rented from private landlord or real estate agent’, ‘rented from employer’, ‘being paid off by you (and/or your partner, i.e. paying off a mortgage)’, ‘being paid off by another family/household member’, ‘owned by you (and/or your partner)’, ‘owned by another family/household member’ or ‘other (please describe)’. These questions were adapted from the Australian Longitudinal Study of Indigenous Children (LSIC). For the purposes of analysis, four tenure types were categorised: public housing, private rental, mortgage (self or family) and ‘other’.

Locational advantage/disadvantage was assessed using an Index of Relative Socio-Economic Advantage and Disadvantage (IRSAD). The index is a summary of socio-economic variables (across income, education, employment, occupation, and housing) and is based on rankings of postcodes across the whole of Australia. The index is split into 10 equal deciles, with the lowest score indicating the most disadvantage and the highest score indicating the most advantage (Australian Bureau of Statistics, 2016). In a separate question, women were asked if they had a health care card (a government concession card provided to families on low incomes). Health care card holders are eligible for subsidised (cheaper) prescription medicines under a pharmaceutical benefits scheme and are eligible for other concessions, including bulk billing of doctors’ appointments, and concessions for utilities, such as electricity and public transport.

Finally, women were asked ‘Are you on the waiting list for:’ ‘a housing trust house in the area where you live now?’, ‘a housing trust house in a different area’, ‘a bigger housing trust house?’, a smaller housing trust house?’ (response categories yes/no).

Financial difficulties

A series of questions on financial hardship were adapted from (LSIC) data (Walter, 2017), Women were asked ‘In the past 12 months, have any of the following things happened to you because you were short of money?’: ‘could not pay bills on time’, ‘could not pay housing payments on time’, ‘went without meals’, ‘unable to heat or cool home’, ‘pawned something/sold something’, ‘assistance from a welfare organisation’, ‘child could not do school activities (e.g. school camp)’ or ‘put off going to the doctor because of the cost’ (response categories yes/no). For analysis, three financial difficulties categories were created; low (0-1), medium (2-3) and high (4-8).

Women’s experiences of partner violence

Women were asked about partner violence with a current partner or former partner in the previous 12 months using the Aboriginal Women’s Experiences of Partner Violence Scale (AEPVS). The AEPVS is a multi-dimensional measure of Aboriginal women’s experiences of physical, emotional and financial partner violence that was co-designed within the Aboriginal Families Study (Glover et al., 2022). In brief, the research team culturally adapted and pre-tested an adapted version of the Composite Abuse Scale (CAS) with Aboriginal women to create a brief culturally-safe self-report measure (Hegarty et al., 2005, 1999). The AEPV scale includes 18 items asking women about different types of behaviours or actions describing physical, emotional and financial abuse (6 items in each domain). Women were asked to report how often they had experienced each behaviour by a current or former partner during the previous 12 months on a four-point scale: ‘never’, ‘once’, ‘several times’ or ‘a lot’ (scored 0-3). Women were identified as experiencing emotional partner violence if they scored ≥3 on the emotional abuse scale (e.g. ‘tried to turn family, friends and children against you); physical partner violence if they score ≥1 on the physical abuse scale (e.g. ‘pushed, grabbed or shoved you’); and financial partner violence if they scored ≥2 on the financial abuse scale (e.g. ‘took money you needed for something else’). For the purposes of this paper, women who reported any type of partner violence were classified as experiencing partner violence.

Analyses

This analysis primarily draws on data collected at wave two. Associations between socio demographic characteristics and housing stability were examined using univariable logistic regression, comparing families with stable housing (reference category) to families with unstable housing. Odds ratios provide the odds of housing stability for each socio demographic characteristic. Multivariable logistic regression was used to examine associations between stable housing and the three primary factors of interest - housing tenure type, financial difficulties and women’s exposure to partner violence. Multivariable models adjusted for mothers’ age for apriori reasons, as several studies show an association between age and housing mobility (Andersen et al., 2017; Biddle & Markham, 2013; Goodman et al., 2018; Skelton, 2002). All analyses were conducted using STATA version 16 (Statcorp, 2019).

Results

At wave 2, 227 mothers and 24 other primary caregivers of the study children completed the questionnaire (72% of the original sample). As questions about housing stability were not included in the carers’ questionnaire, the sample for this paper is restricted to mothers who completed wave 2 follow-up and provided data on how many times they had moved house in the previous 5 years (n = 223, 98.2%).

Participants

The majority of participants were Aboriginal and/or Torres Strait Islanders (91%). Nine percent were non-Aboriginal women who had an Aboriginal and/or Torres Strait Islander child. Most of the Aboriginal women (80%) were connected to a community or language group and reported connections with 35 Aboriginal and Torres Strait Islander language and community groups across Australia.

Sociodemographic characteristics of the sample are reported in Table 1. The mean age of the women was 32.3 years (range 20.6–49.6 years) and the mean age of the study children was 6.4 years (range 5 - 9 years). Slightly more than half of the children were assigned male gender at birth (54.7%) and most children were attending Primary School (96.7%) at the time of follow-up.

Table 1. Sociodemographic characteristics of women participating in the Aboriginal Families Study by housing stability at wave 2 follow-up (n = 223).

		Housing stability (≤2 moves in past 5 years)
	Cohort	Unstable housing	Stable housing
	n (%)	n (%)	n (%)	OR^1 5	[95%CI]
Age of women at follow-up
20-24 years	12 (5.4)	4 (33.3)	8 (66.7)	1.9	[0.5,6.7]
25-29 years	79 (35.4)	38 (48.1)	41 (51.9)	1.0	[Ref]
30-49 years	132 (59.2)	30 (22.7)	102 (77.3)	3.2***	[1.7,5.7]
Relationship status
Not living with partner	108 (48.6)	41 (38.0)	67 (62.0)	1.0	[Ref]
Living with partner	92 (41.4)	22 (23.9)	70 (76.1)	1.9*	[1.1,3.6]
Relationship, not living together	22 (9.9)	9 (40.9)	13 (59.1)	0.9	[0.3,2.3]
Paid employment
No	145 (65.3)	50 (34.5)	95 (65.5)	1.0	[Ref]
Yes	77 (34.7)	21 (27.3)	56 (72.7)	1.4	[0.8,2.6]
Highest level of education
Year 10 or less	114 (52.1)	40 (35.1)	74 (64.9)	1.0	[Ref]
Completed Year 12	33 (15.1)	8 (24.2)	25 (75.8)	1.7	[0.7,4.1]
Diploma/Certificate	57 (26.0)	17 (29.8)	40 (70.2)	1.3	[0.6,2.5]
University Degree	15 (6.8)	6 (40.0)	9 (60.0)	0.8	[0.3,2.4]
Study child’s age
5-6 years	133 (59.6)	39 (29.3)	94 (70.7)	1.0	[Ref]
7-9 years^2	90 (40.4)	33 (36.7)	57 (63.3)	0.7	[0.4,1.3]
Place of residence^3
Metropolitan	100 (45.0)	39 (39.0)	61 (61.0)	1.0	[Ref]
Regional	79 (35.6)	23 (29.1)	56 (70.9)	1.6	[0.8,2.9]
Remote	43 (19.4)	9 (20.9)	34 (79.1)	2.4*	[1.0,5.6]
Number of children living in household biologically related to mother^4
No children	8 (3.6)	5 (62.5)	3 (37.5)	0.2*	[0.0,0.9]
1-2 children	92 (41.3)	35 (38.0)	57 (62.0)	0.6	[0.3,1.0]
3+ children	123 (55.2)	32 (26.0)	91 (74.0)	1.0	[Ref]
Adults in household
Participant/partner	81 (36.7)	18 (22.2)	63 (77.8)	1.0	[Ref]
Participant/partner/other adult(s)	11 (5.0)	4 (36.4)	7 (63.6)	0.5	[0.1,1.9]
Participant	77 (34.8)	29 (37.7)	48 (62.3)	0.5*	[0.2,1.0]
Participant/other adult(s)	52 (23.5)	21 (40.4)	31 (59.6)	0.4*	[0.2,0.9]
Total	223 (100)	72 (32.3)	151 (67.7)

¹ Odd Ratios (odds of stable housing).

² Three study children were 9 years old.

³ Categorised following the Australian Geographic Classification (Australian Bureau of Statistics, 2016).

⁴ Total number of children in the family living with the mother, including the study child.

⁵ *p < 0.05, **p < 0.01, ***p < 0.001.

Housing circumstances

A range of different family household structures were observed, with just over a third of families living in two-adult households comprising the participant and their partner. Another third of households were single adult households. The remaining households were comprised of the participant living with other adults (not partner) – potentially adult children or other family members including extended family. This was also reflected in the range of family members involved in bringing up the study child. Women reported that grandmothers (61.6%), grandfathers (44%), aunts and uncles (57.3%) and older siblings (34.6%) helped them in parenting the study child.

Almost half of the women were living in Adelaide (the major metropolitan city in the state of South Australia) (45.0%), just over a third in regional areas (35.6%) and one in five in remote areas (19.4%). Most were living in a house (91.5%), with a small number living in a unit or flat (5.4%). The majority of women were renting (79.7%), with 49.5% in public housing (government subsidised rental) and 30.2% in a private rental (Table 2). Less than one fifth lived in a house with a mortgage (16.2%). Of these 36 women, 23 (63.9%) were paying off a mortgage, 7 (19.4%) owned their home and 6 (16.7%) reported their home being owned by another family/household member. A small proportion of the sample reported ‘other’ types of housing (e.g. hotel, refuge/shelter and ‘couch surfing’ (4.1%).

Table 2. Social characteristics of women in the Aboriginal Families Study reporting stable housing (≤ 2 moves in last 5 years) at wave 2 follow-up (n = 223).

		Housing stability (≤2 moves in past 5 years)
	Cohort	Unstable housing	Stable housing
	n (%)	n (%)	n (%)	OR^1	[95%CI]
House tenure type
Public Housing	110 (49.5)	26 (23.6)	84 (76.4)	1.0	[Ref]
Private Rental	67 (30.2)	32 (47.8)	35 (52.2)	0.3**	[0.2,0.6]
Mortgage (self or family)	36 (16.2)	5 (13.9)	31 (86.1)	1.9	[0.7,5.4]
Other^2	9 (4.1)	8 (88.9)	1 (11.1)	0.0**	[0.0,0.3]
Level of disadvantage (IRSAD)^3
High disadvantage (1^st/2^nd Decile)	136 (61.3)	45 (33.1)	91 (66.9)	1.0	[Ref]
Medium (3^rd/4^th Decile)	50 (22.5)	16 (32.0)	34 (68.0)	1.1	[0.5,2.1]
Low disadvantage (5^th/10^th Decile)	36 (16.2)	10 (27.8)	26 (72.2)	1.3	[0.6,2.9]
Financial difficulties (past year)
High (4-8)	35 (15.8)	16 (45.7)	19 (54.3)	1.0	[Ref]
Medium (2-3)	53 (24.0)	22 (41.5)	31 (58.5)	1.2	[0.5,2.8]
Low (0-1)	133 (60.2)	33 (24.8)	100 (75.2)	2.5*	[1.2,5.4]
Health care concession card
Yes	172 (77.1)	60 (34.9)	112 (65.1)	1.0	[Ref]
No	51 (22.9)	12 (23.5)	39 (76.5)	1.7	[0.8,3.6]
On a waiting list for public housing
Yes	68 (31.6)	27 (39.7)	41(60.3)	1.0	[Ref]
No	147 (68.4)	43 (29.3)	104 (70.7)	1.6	[0.9,2.9]
Partner Violence (past year)
Yes	82 (38.5)	37 (45.1)	45 (54.9)	1.0	[Ref]
No	131 (61.5)	32 (24.4)	99 (75.6)	2.5**	[1.4,4.6]
Total	223 (100)	72 (32.3)	151 (67.7)

¹ Odd Ratios (odds of stable housing).

² Temporary housing including a refuge, hotel or shelter or ‘couch surfing’.

³ Categorised following the Index of Relative Socio-economic Advantage and Disadvantage (Australian Bureau of Statistics, 2016).

⁴ *p < 0.05, **p < 0.01, ***p < 0.001.

Housing stability

Women’s report of house moves in the last five years revealed one in five had not moved (21.9%), a quarter had moved once (26.9%), 18.8% had moved twice, 13.9% had moved three times, less than ten percent had moved four times (8.5%) and 5.3% moved between five to 16 times. Two thirds of women were categorised as living in stable housing (67.7%). There was no statistical difference between the mean age of women living in stable housing (mean = 33.2, 95%CI 20.6–48.1) compared to women living in unstable housing (mean = 30.6, 95%CI 23.1–49.6).

Socio demographic factors associated with housing stability

Several sociodemographic factors were associated with housing stability (see Table 1). Older women (aged 30 years to 49 years) had higher odds of housing stability compared to women aged 25-29 years (OR = 3.2, 95%CI, 1.7–5.7). Women who were living with a partner had higher odds of housing stability compared to women not living with their partner (OR = 1.9, 95%CI, 1.1–3.6). In contrast, women in single adult households and women living with other adults (not partners) had lower odds of stable housing compared to women living with their partner (OR = 0.5, 95%CI 0.2 − 1.0 and OR = 0.4, 95%CI 0.2–0.9 respectively). Women living in remote areas had higher odds of housing stability compared to women living in a metropolitan location (OR = 2.4, 95%CI 1.0–5.6).

Social factors associated with housing stability

A number of social factors were associated with housing stability, including housing tenure, financial difficulties and exposure to partner violence in the previous 12 months (see Table 2).

Housing tenure

High proportions of women living in public housing and women living in a household with a mortgage reported housing stability (76.4% and 86.1% respectively). In contrast, close to half (47.8%) of the women living in private rental reported unstable housing. Women living in private rental had substantially lower odds of housing stability compared to women living in public housing (OR = 0.3, 95%CI 0.2–0.6). These results were reflected in the number of years women had lived in their current home, with women living in a household with a mortgage reporting an average of 3.3 years, public housing 2.6 years and private rental 1.4 years.

Furthermore, women living in private rental were concentrated in metropolitan Adelaide (50.7%), with just over a third in regional areas (38.8%) and 10.4% in remote areas. In contrast, women living in public housing were more evenly distributed across the locations with 38.2% in metropolitan Adelaide, 33.6% in regional areas and 28.2% in remote areas. Four of the eight women living in ‘other’ types of temporary accommodation were living in metropolitan Adelaide.

Close to one third of the cohort (31.6%) were on a waiting list for public housing, with the majority waiting for a house in a similar area (20.9%), with lower numbers waiting for a house in a different area (10.7%) and looking for a bigger house (7.9%). The largest proportion of women waiting for a house in a similar area lived in a metropolitan area (50%), followed by women living in a regional area (38.6%) and women living in a remote area (11.4%)

Financial difficulties

The majority of women in the cohort had a government concession card available to low-income families, indicating low-income levels across the cohort. Almost half reported that they could not pay bills on time (48.4%), 15.5% went without meals, 15.0% were unable to heat/cool their home, 21.4% of women put off going to the doctor and 10.5% could not pay house payments on time. Women who reported the lowest number of financial difficulties (0-1) had more than twice the odds of living in stable housing (OR = 2.5, 95%CI 1.2–5.4) compared to women who reported the highest number of difficulties (4-8). A relatively high proportion of women living in private rental reported two or more financial difficulties (49.3%), compared to 35.8% of women living in public housing and 27.8% of women living in a household with a mortgage.

Partner violence

The Aboriginal Women’s Experience of Partner Violence Scale (AEPVS) was completed by 213 women (95.5%) of the sample. The majority of women had not experienced partner violence in the previous 12 months. Just over a third of women had experienced physical, emotional or financial partner violence (38.5%) by a current or former partner in the 12 months prior to wave 2 follow-up. Women who had not experienced recent partner violence (past 12 months) had two and a half times the odds of housing stability (OR = 2.5, 95%CI 1.4–4.6) compared to women who had experienced partner violence. Over half of the women who had experienced partner violence reported that they had left the house (59%) or taken their children to stay with family or friends (55%) as ways to keep themselves and their children safe. The proportions of women living in private rental and public housing that had experienced partner violence in the previous 12 months were similar (41.5% and 39.1% respectively). A slightly lower proportion of women living in a household with a mortgage experienced partner violence in the year prior to wave 2 follow-up (25.7%).

Housing tenure, financial difficulties, and partner violence

Multivariable logistic regression analyses were undertaken to obtain a more precise measure of the association between three exposures of main interest (housing tenure type, financial difficulties, and partner violence) and housing stability, taking into account potential confounders (Table 3). All three models adjust for maternal age (included for apriori reasons), and other factors significantly associated with housing stability in univariable analyses (relationship status and place of residence).

Table 3. Multivariable logistic regression models investigating the relationship between housing tenure, financial difficulties, partner violence and housing stability (n = 223).

	MODEL 1 Housing Tenure	Model 2 Housing Tenure & Financial difficulties	MODEL 3 Housing Tenure & partner violence
	Adj. OR^1 [95%CI]	Adj. OR [95%CI]	Adj. OR [95%CI]
Housing Tenure Type
Public housing	1.0 [Ref]	1.0 [Ref]	1.0 [Ref]
Private Rental	0.3**[0.2,0.7]	0.4**[0.2,0.8]	0.4**[0.2,0.7]
Mortgage (self or family)	1.2 [0.4,3.9]	1.2 [0.4,3.6]	1.3 [0.4,4.0]
Other^2	0.0**[0.0,0.4]	0.1**[0.0,0.5]	0.1*[0.0,0.5]
Financial Difficulties (past year)
High (4-8)		1.0 [Ref]
Medium (2-3)		1.2 [0.4,3.1]
Low (0-1)		2.2 [0.9,5.4]
Partner Violence (past year)
Yes			1.0 [Ref]
No			2.2*[1.1,4.3]
Mother’s Age
20-24 years	1.1 [0.3,4.1]	1.2 [0.3,4.9]	1.0[0.3,3.9]
25-29 years	1.0 [Ref]	1.0 [Ref]	1.0 [Ref]
30-49 years	2.4**[1.2,4.7]	2.8**[1.4,5.6]	2.6**[1.3,5.1]
Relationship Status
Not living with partner	1.0 [Ref]	1.0 [Ref]	1.0 [Ref]
Living with partner	1.8 [0.9,3.5]	1.7 [0.8,3.3]	1.4 [0.7,2.8]
Place of Residence
Metropolitan Area	1.0 [Ref]	1.0 [Ref]	1.0 [Ref]
Regional	1.7 [0.8,3.4]	1.5 [0.7,3.1]	1.5 [0.7,3.2]
Remote	1.7 [0.7,4.4]	1.5 [0.6,3.9]	1.6 [0.6,4.1]
Total

¹ Odds of stable housing for women in different housing tenure compared to women living in government subsidised public housing.

² Temporary housing including a refuge, hotel or shelter or ‘couch surfing’.

³ Categorised following the Australian Geographic Classification.

⁴ *p < 0.05, ** p < 0.01, ***p < 0.001.

Housing tenure

After adjusting for socio demographic factors, women living in private rental had lower odds of housing stability compared to women living in public housing (Model 1: Adj.OR = 0.3, 95%CI 0.2–0.7). The odds of housing stability remained lower for women living in private rental accommodation even after adjusting for financial difficulties (Model 2: Adj.OR = 0.4, 95%CI 0.2–0.8) and partner violence (Model 3: Adj.OR = 0.4, 95%CI 0.2–0.7).

Financial difficulties

After adjusting for housing tenure and sociodemographic factors, women who reported low levels of financial difficulties remained at higher odds of housing stability compared to women who reported high financial difficulties (Model 2: Adj.OR = 2.2, 95%CI 0.9–5.4), with the confidence interval indicating borderline significance.

Partner violence

After adjusting for housing tenure type and socio demographic factors, women who had not experienced partner violence in the previous 12 months had higher odds of housing stability compared to women who experienced partner violence (Model 3: Adj.OR = 2.2, 95%CI 1.1–4.3).

Of interest is that women’s age was the only socio demographic factor that remained significantly associated with housing stability in the multivariable models, with women aged 30 years or older more likely to have housing stability.

Discussion

We found a mix of household family structures across the cohort, with a high proportion of female headed households. A range of family members were involved in bringing up the study children, including partners, grandparents, aunties, uncles and siblings.

The majority of families (67%) lived in stable housing (0-2 moves) in the 5 years prior to wave 2 follow-up when the children were aged 5-9 years. Our findings demonstrate house tenure, level of financial difficulties and exposure to partner violence played a significant role in housing stability. Older maternal age and living with a partner were both key factors in stable housing. Both relate to financial position. Studies have shown older women have had more time to save for a rental or mortgage deposit, time to study and build wealth as security from hardships, such as sickness, a relationship breakdown, eviction or job loss (Cripps & Habibis, 2019; Mallett et al., 2011; Phinney, 2013). We found women who reported the lowest levels of financial difficulties had higher odds of housing stability. Conversely, close to half of the women who reported a high level of financial difficulties had unstable housing. It is widely acknowledged that low income households move more frequently (Cohen & Wardrip, 2011; Hastings, 2023; Lawrence et al., 2016; Phinney, 2013) and are more likely to move to disadvantaged areas (Baker et al., 2016). Consistent with this research, over half of the cohort lived in the lowest two decile areas as ranked by the Australian Bureau of Statistics in terms of most disadvantaged/least advantaged post codes. Although we did not ask women about income levels, most women in the cohort held a government provided health care concession card indicating low-income levels across the cohort. Low income is a risk factor for housing instability.

House tenure type was associated with housing stability. Women living in a house with a mortgage and those living in public housing reported fewer financial difficulties and had lived in their homes for the longest period (median 3.3 and 2.6 years respectively). Women living in a house with a mortgage had the highest levels of stability. However, it is evident that a mortgage is not an affordable option for many families, with under one-fifth (16.2%) of the cohort living in this tenure type. This is consistent with other literature showing fewer Aboriginal families live in a house with a mortgage. In a study on urban households with Aboriginal children in Sydney Australia, 19% of families lived in a house with a mortgage, 60% in public housing and 21% in private rental (Andersen et al., 2017).

Women living in private rental (median 1.4 years) had significantly lower odds of housing stability and reported increased financial difficulties. This is despite rental subsidies in Australia being available for low-income households living in private rental.⁴ The increased financial difficulties reported indicate rental stress⁵ and suggests that government subsidies are not providing enough financial support. In 2022, over one third (36%) of Aboriginal recipients of Commonwealth Rent Assistance (CRA) in Australia were still in rental stress (Australian Institute of Health and Welfare, 2023). The proportion of Aboriginal CRA recipients in rental stress has increased over time, from (30%) in 2013 (Australian Institute of Health and Welfare, 2023).

A 2021 Australian report found housing tenure to be the factor with the largest impact on the decision to move versus stay (James et al., 2021). Renters were three times as likely to move as home owners, possibility reflecting the high costs of home purchase and lack of tenure security in the private rental market (James et al., 2021). Private rental leases are on average 6 to 12 months (Moskos et al., 2022) and Aboriginal families face the added barrier of discrimination and racism (Andersen et al., 2018).

The higher number of house moves and financial difficulties experienced by Aboriginal families living in private rental indicate a need for policy change. Our findings build on existing policy recommendations and calls for: a review of rental subsidies to meet the needs of low income families living in private rental (Blunden & Flanagan, 2021): advocacy for rental terminations affecting families with children; (Valentine et al., 2020) and advocacy to address discrimination and short-term leases in private rental (Valentine et al., 2020). Increasing and supporting the number of Aboriginal landlord organisations would provide families with more housing choices that recognise and value the differing cultural needs of Aboriginal families, for example homes designed for larger families and visiting extended family. Flexible tenancy arrangements could take into consideration cultural obligations and extended family responsibilities (Valentine et al., 2020).

Close to one third of women in the cohort were on a waiting list for public housing. Half of these women lived in a metropolitan area and over one third lived in a regional area. The high proportion of women in private rental within metropolitan areas suggests there are fewer public housing options available. Most women reported they wanted a house within the area they currently live, indicating established family, friendship and community networks. More public housing options in these local areas would support more families to have stable housing and enable mothers and children to stay close to support networks. Stable housing could mean more children stay at the same childcare, kindergarten or school and families have access to continuous and ongoing health care services.

The number of house moves across the overall cohort was relatively high. Unfortunately, we did not ask families why they were moving. Research on mobility has found Aboriginal people move more due to cultural factors; maintaining connections with family and kin, connecting to Country, or to access cultural activities (Anderson & Collins, 2014; Gee et al., 2014; Memmott et al., 2006; Tedmanson et al., 2022). These factors are associated with health benefits for many Aboriginal people (Gee et al., 2014; Verbunt et al., 2021). Temporary moves may also occur for sorry business such as attending funerals, or for families to access health services in metropolitan areas (Habibis et al., 2011). Some of these reasons for moving were evident at data collection, with families moving between regional and urban areas, to be close to other family members and to stay connected to Country. Fieldwork team members noted an increase in families moving during the period of the COVID pandemic, with families moving back to Country or moving closer to family.

High mobility in Aboriginal families combined with housing disadvantage is consistent with findings from other studies with Aboriginal families (Andersen et al., 2017; Australian Government Department of Social Services, 2015; Dockery et al., 2013; Lowell et al., 2018; Nikolof et al., 2023). Globally, Indigenous people are increasingly moving to metropolitan areas and experiencing higher levels of mobility compared to families living in remote areas (Andersen et al., 2018; Anderson & Collins, 2014; Snyder & Wilson, 2012, 2015; Tedmanson et al., 2022). Findings from the Western Australian Aboriginal Child Health Survey (WAACHS) showed Aboriginal families living in regional and metropolitan areas had higher mobility (Australian Government Department of Social Services, 2015). Recent studies provide evidence of unwanted mobility, with mothers or caregivers reporting being forced to move in the last 12 months; 12% of caregivers from a longitudinal cohort study of Aboriginal children in New South Wales (Andersen et al., 2017) and 9.2% of caregivers in a cross-sectional survey in Western Australia (Australian Government Department of Social Services, 2015). Forced residential mobility occurs when households wish to remain in a given location, however find themselves forced to move in response to things outside their control, for example lease arrangements, unaffordable or poor quality housing (Baker et al., 2019). A common driver of forced mobility is women who are experiencing partner violence.

Women in the study who reported experiencing partner violence in the past year, had over double the odds of living in unstable housing. Over half of the women who had experienced violence said they had left the house or took the kids to stay with family and friends to protect themselves. A report on Improving Housing and Service Responses found the three most dominant pathways in the aftermath of family violence were: staying with family/friends, staying at a refuge/safe house or remaining in an unsafe home (Cripps & Habibis, 2019). Acute shortages in crisis, transitional and long-term housing mean that Aboriginal women and children cycle through one of these three options repeatedly. Access to crisis and transitional housing can be delayed for women experiencing financial partner violence. Being unable to access money, bankcards or documents to confirm identity can result in significant delays for women, who often need to re-establish their identity to access government income support in order to be placed in short-term housing (Cripps & Habibis, 2019). Family violence related housing debt and poor rental records create further barriers for women in accessing private rental and public housing (Valentine et al., 2020). The Improving Housing and Services Responses report also found that women who sought long-term public housing separate to their partner were routinely waiting in excess of a year (Cripps & Habibis, 2019).

Moving between transitional housing and delays to long term housing can mean women are at increased risk of losing custody of their child/ren. Both housing instability and children’s exposure to family violence are risk factors for child protection⁶ service involvement (Valentine et al., 2020). In these circumstances women are faced with impossible choices, either moving multiple times between family, crisis or transitional housing or returning to an unsafe home (Cripps & Habibis, 2019). Aboriginal children are overwhelmingly over-represented in child removals in Australia (Australian Institute of Family Studies, 2020). There is an urgent need to improve service coordination between housing, child protection and welfare services.

The majority of women experiencing partner violence reported multiple types of violence, with few women reporting emotional, financial or physical abuse alone. Findings showed over one third of the cohort experienced physical, emotional or financial partner violence. These findings provide important evidence of the high numbers of women experiencing partner violence and the significant associated risks women face in maintaining access to safe and stable housing.

The high number of women experiencing partner violence is reflected in the growing number of Aboriginal women and children in Australia accessing homelessness services. In 2022, over half of the Aboriginal clients receiving support from Specialist Homelessness Services were women (62%). Close to one third of Aboriginal clients were living as a single parent with children(ren) (34%) and one in three clients had experienced family violence (36%) (Australian Institute of Health and Welfare, 2022). Considering there are clear intersections between housing, legal and welfare services, it is essential that services and agencies work alongside families and work in coordination. Integrated service responses to family violence are considered the ‘gold standard’ in Australia and internationally (Cripps, 2023). An example of a model involving collaboration between agencies and Indigenous families can be found in Waikato and Christchurch in New Zealand. The Integrated Safety Response to Family Violence Model (ISR) was designed to look after the safety and wellbeing of women and children affected by family or sexual violence. The ISR is a multiagency intervention that brings together the efforts of several groups and agencies – including but not limited to housing, the police, Ministry for Children (known as Oranga Tamariki), Corrections, Justice, Social Development and Education; and other specialist family-violence non-government organisations and Maori services in the communities (New Zealand Government, 2023).

Strengths and limitations

The Aboriginal Families Study is the only prospective cohort of Aboriginal children, their mothers and caregivers born in South Australia and guided by an Aboriginal Governance Group throughout the study. We have used internal cohort comparisons and comparisons with Indigenous population groups. The analysis has taken a strengths-based approach focusing on stability with the aim of recognising and supporting healthy and strong Aboriginal families, communities and culture into future generations (The Lowitja Institute, 2020).

Women’s experience of physical, emotional and financial partner violence was reported using a new community co-designed measure, the Aboriginal Women’s Experience of Partner Violence Scale (AEPVS) (Glover et al., 2022). This is one of the first few studies to have employed a culturally safe, self-report measure of Aboriginal women’s experiences of partner violence.

It is possible that the families who moved the most did not participate in the wave 2 follow-up (72% of the original sample), due to the difficulty of locating and contacting families with high housing mobility. As a result, the study may underestimate the extent of housing mobility among Aboriginal families in South Australia with primary school age children. Reasons for non-participation in wave 2 follow up were: mother passed away (n = 3), study child passed away (n = 2), mother withdrew (n = 23), mother/primary carer of the study child unable to be contacted (n = 56). We are also mindful that the sample was restricted to the mothers of the study children taking part in wave 2 follow-up, due to questions about housing stability not being included in the questionnaire completed by other primary caregivers. We are aware that fathers and other primary caregivers play an important role in families and in raising children. Aboriginal families can involve a wider group of family members with shared intergenerational responsibilities in raising children. Children may spend time in other households and have high levels of interaction with extended family or kin (Walter, 2017). It is likely that this may have resulted in some under-identification in children moving house, either between parents’ homes, between family members or for children living with other care givers.

Families were not asked why they moved house. Further research is needed to explore why families are moving and if children are moving between schools. Cultural factors such as staying connected to family, kin, and Country are important drivers of mobility and should be considered (Anderson & Collins, 2014; Habibis et al., 2011; Tedmanson et al., 2022). Future research is needed to examine the impact of mobility on the health and wellbeing of adults and the health, developmental and learning outcomes for children. It is important to factor in broader cultural determinants of health including positive and/or negative outcomes to health and wellbeing. Prioritising the involvement, collaboration, and empowerment of Aboriginal people and leadership is critical for future research (Barnabe, 2021).

Conclusion

Our findings show that tenure type, financial difficulties and experience of partner violence are key factors associated with housing stability for Aboriginal families. While the majority of the cohort had stable housing, one third of the cohort were experiencing housing instability. Understanding factors associated with housing stability/instability can inform services and policy stakeholders where to target support.

There is a clear need for support from multiple sectors, including housing, law, health and welfare. Support from these sectors need to integrate policy responses and work in collaboration to provide support for women who may be experiencing a combination of factors; limited housing options, financial difficulties and experience of partner violence. Delays for women in accessing crisis and transitional public housing in the aftermath of family violence places women and children at risk of further violence, homelessness and child removal. There is also an urgent need to address delays for long-term public housing.

Our findings build on existing calls for increased government investment in public housing stock and low-cost affordable housing options to support families with young children in metropolitan and regional areas. Prioritising housing support for younger women and children would reduce risk factors associated with housing instability and lead to improved health and wellbeing of parents and children.

Acknowledgments

The authors acknowledge the Aboriginal and Torres Strait Islander Custodians of the Lands and Waters of Australia. We would like to thank members of the Aboriginal Families Study Aboriginal Governance Group for their support, input, and guidance throughout the research process, including in the development of the study protocol, review of study findings, and interpretation of results. We would also like to acknowledge the support provided by the Aboriginal Families Study Writing Group.

Disclosure statement

The authors declare no conflict of interest.

Additional information

Funding

The Aboriginal Families Study was supported by two Australian National Health and Medical Research Council (NHMRC) project grants (#1004395 and #1105561) and by SA Health funding awarded to the study. A.N. is supported by an Australian Government Research Training Program Scholarship and the top-up scholarship supported by the Royal Children’s Hospital Research Foundation and Murdoch Children’s Research Institute. Stephanie Brown is supported by an (NHMRC) Leadership (L2) Investigator Grant (GNT2008925). Yvonne Clark by is supported by NHMRC Emerging Leadership (EL1) Grant (GNT1178590). Research conducted at the Murdoch Children’s Research Institute is supported by the Victorian Government’s Operational Infrastructure Support Program.

Notes on contributors

Arwen Nikolof

Arwen Nikolof is a Narungga/Kaurna woman from South Australia. She is a PhD candidate at the University of Melbourne and Murdoch Children’s Research Institute. She has worked in Aboriginal health for over twenty years, in policy, planning and service development within community services, government and the research sector.

Stephanie J. Brown

Stephanie J. Brown is a social epidemiologist and health services researcher whose work focuses on improving the health and wellbeing of children, young people and families and reducing social inequity. She established and leads the Intergenerational Health group at Murdoch Children’s Research Institute, and is also Co-lead of the Institute’s Aboriginal Health Program. Her team are working with Aboriginal communities, refugee communities and women and children exposed to family violence to address gaps in evidence regarding effective intervention strategies to promote resilience and enable healing and recovery from the effects of trauma and family violence.

Yvonne Clark

Yvonne Clark is a Kokatha/Wirangu women from South Australia. She a research associate professor in a joint appointment with the South Australian Health and Medical Research Institute and the University of SA. She is a psychologist and works on topics concerning wellbeing, trauma and lateral violence in relation to Aboriginal and Torres Strait Islander families, children and young people.

Cathy Leane

Karen Glover is a Mein:tnk and Wotjobaluk women from South Eastern South Australia and north western Victoria. Karen has worked for over 30 years in the government and non-government Aboriginal health and community service sectors, including in policy, planning, service development, management and advocacy. Karen holds a joint appointment as Senior Research Fellow at the South Australian Health and Medical Research Institute and the Murdoch Children’s Research Institute, where she is Co-leader of Aboriginal health within the Healthy Mother Healthy Families Group.

Karen Glover

Cathy Leane is a Dharug woman. She has more than twenty years of experience working in early childhood services. Previously she was Senior Policy Advisor for the South Australian Department of Premier and Cabinet. Her expertise includes policy development, implementation, and research with a whole of government perspective. Her current role with Women’s and Childrens Health Network includes coordination of research partnerships, research translation and community engagement.

Deirdre Gartland

Deirdre Gartland leads a program of research in resilience, mental health and measurement development, in particular for families facing social adversity and inequity. A primary research focus has been investigating the impacts of partner violence, and how to better support women and children. Her research in the last 7 years has had a particular focus on child resilience, including factors that protect women and children’s mental health, particularly for Aboriginal families, refugee-background families and families exposed to violence.

Notes

1 The term Aboriginal used throughout this paper is intended to refer to Aboriginal and Torres Strait Islander peoples in Australia. We acknowledge the vast and rich cultural diversity that exists within and between these communities.

2 To be connected to Country refers to an area to which people have a traditional or spiritual association, a sense of connection or belonging.

3 State government agency now called South Australian Housing Authority.

4 Families in Australia who rent in private or community housing are eligible for the Commonwealth Rent Assistance (CRA), a capped cash payment to households in receipt of income support.

5 ‘Rental stress’ is defined by the Australian government as spending more than 30% of gross income on rent.

6 Government agency responsible for overseeing child protection with the power to remove children from families.