• Open Access

Causes of homelessness among older people in Melbourne, Australia


Correspondence to: Dr Alice Rota-Bartelink, Wintringham, PO Box 193, Flemington, Victoria 3031. Fax: (03) 9376 8138; e-mail: alicerota@wintringham.org.au


Objective: A comparative study of the causes of new episodes of homelessness among people aged 50 years and over has been undertaken in Australia, the United States and England. Semi-structured questionnaires were used to collect information on the circumstances and problems that contributed to homelessness.

Methods: This paper presents the findings from Australia, where information was obtained from 125 older homeless people (aged 50+ years) and their key workers in Melbourne. All three participating nations followed identical research methodologies.

Results: The factors most frequently reported by respondents as contributing to their homelessness were problems with people with whom they lived, followed by physical and mental ill-health and problems associated with the housing itself. The most frequently reported factors by case workers were problems with alcohol, followed by physical and mental health factors.

Conclusions: This study demonstrates a significant under-utilisation of housing and support services among recently homeless older people and provides evidence that people who had previously been homeless appear to be more resigned to their homelessness than do those who had not experienced homelessness before. Significant issues relating to depression and gambling were also noted. The findings support the need for more targeted, specialised services to be developed or improved such that older homeless people can readily gain access to them and for improved collaboration or information exchange among housing providers and welfare agencies.

Despite the richness of services available to older Australians, there are still people in the community who are not accessing or are under-utilising these services. For those who are at risk of homelessness and fortunate enough to successfully engage with an appropriate housing provider, a critical shortage of suitable, affordable housing places further obstacles in the path to securing a home. The 2001 Census and other sources identified 99,000 homeless people, of whom 24,227 (24%) were aged 45 years or more.1 It has been estimated that there may be in excess of 250,000 older people at risk of homelessness because of increases in the number of older people on fixed incomes and those relying on insecure housing.2 A recent Government-funded review of aged care services observed that the elderly homeless population, albeit small, is one of the most difficult groups to place in residential care.3 This can be attributed to several factors.

Homeless people often have poor interpersonal skills and are suspicious of people they do not know, including service providers, and it takes a great deal of time to build a relationship of trust.4 It has been demonstrated that homeless populations have a higher rate of serious morbidity and premature mortality compared with the general population, with Westernised countries reporting an average age of death between 42 and 52 years.5,6 There is increasing acknowledgement that the minimum age of eligibility for homeless people to gain access to aged care services should be 50 years of age as opposed to 65.5,7 The strict eligibility criteria employed by many health and welfare agencies deny prematurely aged older homeless people access to their services.

Historically in Australia, support for the elderly homeless has been managed, with varying degrees of success, in an unco-ordinated, ad hoc manner by a range of service systems. Complex funding processes mean that resources to support elderly homeless are spread over multiple Commonwealth and State and Territory government portfolios, non-government organisations and service providers, creating practical obstacles to agencies comprehensively addressing an individual's complex needs.8 Australia-wide trends include declining levels of home ownership because of such factors as the gentrification of low-cost housing stock and the growing number of people renting rather than buying. These have created obstacles for low income people to access the private market through lack of supply, higher-income households occupying low-cost housing, lack of affordability, discrimination, lack of legislative protection against unfair rent increases, lack of legislation to ensure minimum housing standards, and the operation of unregulated tenancy databases.9

Since the late 1980s, increased attention has been given to the problems of older homeless people,10-13 many of whom were homeless for the first time, raising questions about the reasons why older people become homeless and whether their homelessness can be prevented. Two major reports have examined the plight of older homeless Australians. Judd et al. (2004) presented evidence that unexpected life events such as the death of a spouse, relationship crises, decline in health, or rental increases commonly led to accommodation crises. A shortage of public and community housing options was attributed to a severe restriction in the availability of affordable and adequate housing.14 Based on data obtained from the Supported Accommodation Assistance Program (SAAP), Lai (2003) found that the main reasons for older men to seek assistance were financial difficulties, substance abuse, and/or having recently arrived in the area without supports. Older women cited issues such as domestic violence, financial difficulties, having recently moved into the area with minimal support, and psychiatric illness as their reasons for seeking assistance. Older clients were more likely than younger clients to have lived in their own homes before seeking SAAP assistance.15

Study objectives

With the aim of increasing the understanding of the reasons for homelessness among people aged 50 years and over, and contributing to the development of prevention practice, a collaborative and comparative study between three nations (England, Australia and the US) was undertaken. Led by the Sheffield Institute for Studies on Ageing (SISA), the Melbourne-based Wintringham, funded by the Commonwealth Department of Family and Community Services, was the Australian partner in the study. Wintringham is a non-profit welfare company that provides safe, secure, affordable, long-term accommodation and high-quality services to 750 frail elderly men and women who are homeless or living in marginalised accommodation. The conceptual model underpinning this study conceives homelessness as a function of structural factors and policy gaps, personal problems and incapacities, and welfare-service delivery problems.16


A full account of the genesis, aims, design and execution of the three-country study has previously been published.16 Identical methodologies including selection criteria, questionnaires and data recording processes were employed in all three participating countries and were subject to regular quality and consistency audits throughout the study. The study population included people who had become homeless during the two years prior to participation and were aged 50 years or over when they became homeless. People who had previously been homeless were included if they had been housed for at least 12 months prior to their current episode of homelessness. The focus on people who had recently become homeless was intended to enhance the reliability of the information about the circumstances that led to homelessness and to overcome problems of memory. In Melbourne, 125 people were interviewed. The sample principally drew on Wintringham's clients, but other organisations were approached to increase participant numbers, particularly of women. Outreach service staff played a crucial role in establishing initial contact with potential participants for the study.

Homelessness was defined as: (i) those without conventional accommodation; (ii) people living in accommodation that was intended as only temporary; (iii) those who were ‘doubled up’ with relatives, friends or other households; (iv) those without housing after eviction, discharge from prison or hospital, etc; and (v) people who had housing but were unable or refused to stay in the accommodation.

A semi-structured questionnaire was used to collect information on respondents’ circumstances and problems prior to homelessness. The interviews (completed in 2003) were conducted face-to-face by experienced researchers and focused on: (i) background details; (ii) services received; (iii) service responses; and (iv) perceived reasons for becoming homeless. As a measure of reliability on the self-reported information, interviews were also conducted with each individual respondent's key worker (or case worker) to assess their understanding of the events and states that led to the person becoming homeless.

Non parametric analyses (Pearson's r correlations) and F distributions were principally used to analyse the data, with the level of significance (alpha) set at 0.05. Cross-national findings of this study have been previously published.16-19 We will now present the results from an Australian perspective.


Of the 125 people interviewed, the gender split was 93 male (74%) and 32 female (26%). Almost half of the respondents were aged in their fifties, with 36% in their sixties and 16% in their seventies. Some 37% of the respondents were born outside Australia and approximately one-quarter were born in a non-English speaking country. Twenty-six per cent of the respondents had served in the armed forces (more than twice the expected proportion based on Wintringham's experience and previous research20). A significant proportion of men lived on their own immediately prior to becoming homeless compared with women (F=10.12, df= 1, 123, p=0.002). Twelve respondents first became homeless before 30 years of age and a further nine people before the age of 50. Almost three-quarters (70%) of those aged 60 and over had never experienced a prior episode of homelessness. More than half of the respondents reported that they had lived in their former home for three years or more prior to becoming homeless. The study population demographics, compared with the English study, are displayed in Table 1.

Table 1.  Study population demographics.
CharacteristicsMen %Women %NumberTotal
    % Aust% UK
  1. Notes:

  2. (a) Current episode of homelessness.

  3. (b) Never married.

  4. (c) Includes cohabitating.

Age became homeless (years)a     
Marital status     
Previously homeless     
Housing history     
 Self-contained house/apartment65.690.69072.074
 Room in shared house10.83.1118.814
 Rooming house/SRO18.301713.65
Attached to job1.13.121.67
Employment history     
 Mostly employed63.443.87358.471
 Employed intermittently28.025.03427.222
 Mostly unemployed8.631.31814.47
Mental health problems (self reported)     
 Depression/feeling low54.878.17660.858
 Any other mental health problems35.450.04939.261
  Other problems6.43.175.6 
Physical health problems     
 Reported by subjects61.362.57761.670
 Evident or suspected by staff75.371.99374.458
Alcohol problems     
 Reported by subjects43.228.15443.236
 Evident or suspected by staff77.543.88164.843
Drug misuse problems     
 Reported by subjects12.912.51612.86
 Evident or suspected by staff11.825.01915.23
Gambling problems     
 Reported by subjects37.612.63628.85
 Evident or suspected by staff53.86.25241.65
Number of subjects9332125 132

Figure 1 displays the frequency with which various factors were reported by respondents as having contributed significantly to their current state of homelessness and the influence of gender. The factors most frequently reported were problems with other people with whom they lived, followed by physical health and problems with the housing itself. The most frequently reported factors by case workers were problems with alcohol, followed by physical and then mental health factors. The majority (86%) of respondents were in contact with relatives or friends before becoming homeless, but only 49% received help from these contacts. However, by far the most commonly cited contact or friendship was with fellow homeless persons. Women were more likely than men to seek informal help from friends and relatives. Just over half of respondents received no formal assistance from professional services prior to becoming homeless.

Figure 1.

Percentage frequency of reported reasons for leaving last place of residence versus gender (based on respondents' reports).

Note: Category ‘Other problems’ includes issues relating to physical and mental health.

The analysis draws attention to the comparison of responses from those respondents who became homeless for the first time in later life (60%) with those who had experienced homelessness on one or more previous occasions (40% of the sample). Comparison of these two subgroups confirms two distinct later life pathways. Men were more likely than women to have previously experienced an episode of homelessness and significantly more likely to have been homeless for periods totalling longer than three years (F=4.1, df=1,123, p=0.045). Those who had previous experience of homelessness were twice as likely than those who first became homeless to have never been married (42% compared with 23%) and the men were less likely to have worked regularly throughout their adult lives (50% compared with 74%). Those with previous experience of homelessness were also more likely to have moved several times in the three years prior to becoming homeless. Almost all (92%) of the first-time homeless had only one or two homes during this period, compared with 68% of those with prior experience of homelessness.

Those who had previously experienced homelessness were also twice as likely as those who had never been homeless to report problems with alcohol (60% compared with 32%) and more likely to report having both mental health and alcohol problems (44% compared with 24%) and occasionally using illegal drugs (24% compared with 5%). The previously homeless were less likely to receive assistance from friends or relatives prior to the last episode of homelessness (40% compared with 55%), more likely to attend a day centre for meals and less likely to access a housing support worker or social worker to assist them.

The experience of primary homelessness is frequently associated with premature ageing, leading to premature admission into residential care, extended stays in hospital, and the requirement of a higher level of support.21 It was found that almost half the Australian respondents (45% or 56 respondents) reported that their main source of medical assistance was received from the local public hospital. Seventy-eight per cent of all respondents reported having health problems prior to becoming homeless, of which one-third (30%) stated that their physical health problems contributed significantly to their homelessness. Musculo-skeletal and cardiovascular problems were the most commonly reported medical complaints (see Table 2).

Table 2.  Physical health problems prior to homelessness (respondents’ reports).
health problems
Neurological disorders20.415.62419.2
Endocrine disorders11.821.91814.4
Digestive system12.99.41512.0

Sixty-one per cent of the respondents in this study reported having felt depressed or low prior to becoming homeless and a further 28% reported having other mental health problems either in conjunction with their depression or in isolation. The presence of mental health problems was statistically significant (p<0.001). Men were significantly more likely than women to suffer depression immediately prior to becoming homeless (Pearson=50.2, df=1, p<0.001). To compound this issue, only 56% of men who reported that they had mental health problems had received treatment for it as compared with 78% of women. People aged 50-59 years were more likely than those aged 60+ to report depression, while those who had previously been homeless were more likely than those who had never been homeless to report being depressed (69% compared with 57%). One-quarter of the entire population reported that their depression or mental health problems had contributed to their homelessness. Case worker reports demonstrated that there was a significant under-reporting of depression (84% case worker reported versus 61% self-reported), particularly among men (84% versus 55%).

Seventy-seven per cent of men and 44% of women were identified as having alcohol problems. Two-thirds (67%) of the respondents who reported having issues with excessive alcohol consumption had received assistance for their alcohol problems from self-help groups such as AA or professional counselling. Contrary to the culture surrounding youth homelessness, the aged homeless respondents in this study reported only a small incidence (13%) of problems with illicit drugs. However, this may increase as survivors of heavy drug use are themselves ageing and presenting with a number of drug-related ailments. All respondents who reported drug issues were aged between 50-59 years, most of whom had not sought professional assistance for their addiction.

Men were significantly more likely to report having problems with alcohol and gambling. Further analysis showed that men (46%) were more likely than women (16%) to self-report their gambling. A Victorian Casino and Gaming Authority report (1997) showed that older men were also more likely to spend more of their income on gambling each week ($25 compared with $14 for women).22 However, a significant relationship was found between those living alone and gambling (46%) compared with those who lived with others (28%). It was striking to note that despite the public anti-gambling campaigns and the numerous support agencies available to assist problem gamblers, 85% of those who reported gambling issues did not seek assistance for their problem.


It is well known that many people in need of support find it difficult to access appropriate services due in part to the complexity arising from the division of powers in the health and welfare systems.23 The study provides new data to illustrate the inadequacies of the public service system in meeting the needs of older people who are at risk of becoming homeless. The value of a comparative perspective between countries lies in the leverage it gives to identifying features of the welfare system that work to lessen or increase the risk of homelessness among this group. The same goes for broad cultural differences.

The English partners summarised their findings by concluding that: “For most respondents, a combination of vulnerabilities and negative events accumulated over time and led to a progressive increase of housing instability”. They identified five “packages of reasons” that created distinctive pathways into homelessness. These were: (1) mobility or functioning difficulties that derived from physical and mental health problems; (2) financial problems and rent arrears; (3) the death of a relative or close friend; (4) the breakdown of a marital or cohabiting relationship; and (5) disputes with co-tenants and neighbours.17 These pathways do not differ significantly from the Australian findings; however, an additional factor related to problems with housing could be added to this list based on our results (see Figure 1). Many of our respondents reported that issues such as large increases in rental payments, unacceptable or unsafe states of disrepair, and the closure of rooming houses and hotels had contributed significantly to their homelessness.

Compared with the findings from England on the prevalence of factors contributing to homelessness, the most aberrant result is seen in the high ranking position of gambling problems in Melbourne. Some 38% of respondents in the Australian sample self-identified as having gambling problems (case workers believed the proportion to be considerably higher); in England this was true of only 5% of respondents.18 There were also notable differences in the rates of reported depression among the Australian respondents compared with the English findings, which is in alignment with recent action to make depression a national health target priority.

A large comparative study into youth homelessness highlights the differences in antecedents to homelessness that can be attributed to age. Project i was conducted in Melbourne and Los Angeles between 2000 and 2005. Among a population of 12 to 20 year-olds, family conflict was reported as the most common reason for leaving home, followed by domestic violence, personal/parental drug and alcohol abuse, personal anxiety and depression, and a desire for adventure/independence.24

A lack of access to mainstream welfare services has been highlighted in previous Australian studies25,26 and was confirmed in the findings of this study (although lack of a comparison group reduces the strength of these findings). While the reasons for the under-utilisation of services are complex, it is difficult to escape the conclusion that people who had previously been homeless appeared to be more resigned to their homelessness than those who had no previous experience. People who were recently homeless appeared to be more willing to access or accept formal and informal supports than people whose prior experience had resulted in an impression that these services were frequently unwelcoming and inappropriate.

However tempting, it is difficult to draw conclusions from the frequencies with which each of these factors were nominated by respondents as having influenced their homelessness; the factors cannot be considered independently. The effects of one factor can influence another. For example, a change in physical health (health problems) could result in mobility problems that impede access and ability to maintain accommodation (housing problems). These major life changes could exacerbate depression (mental health problems), placing strain on a marriage (disputes with co-tenants), which may result in a breakdown of the relationship. Any one of these factors could be nominated as a principal cause of their homelessness. The factor to which they assign the greatest impact may be influenced by such determinants as that which may have been more recent or that which may have elicited a greater emotional response. The selection is therefore is entirely subjective.

The difficulty encountered when trying to evaluate self-reported information from a population of older homeless people is that it relies on responses given by people with a high likelihood of having mental or psychiatric deficits resulting from alcohol-related brain impairment and dementia. Additionally, these people have a sense of pride that sometimes finds expression in an overly positive view of their previous life events. Many of our respondents demonstrated a lack of insight, which may have been a complicating factor in our attempt to elicit reliable responses. This study utilised key worker feedback responses to control for potentially unreliable participant responses. Through this method, it was noted that factors such as excessive alcohol consumption and problematic gambling were often under-reported by respondents and factors such as previous home ownership, financial security and military service were often over-reported in comparison with Wintringham's practical experience and Australian research findings.14,20,26


The outcomes of this study may in some part dispel the stereotype of lifetime homelessness ultimately leading to homelessness among older Australians. Due to the nature of the selection criteria, the results of this study would be expected to demonstrate a higher proportion of first-time homeless; however, this does not negate the fact that there is still a significant proportion of newly homeless older people who through circumstances such as significant health changes, relationship crises and loss of accommodation are left with insufficient resources to maintain independent housing. This study has demonstrated that the longer a person has been homeless, the less likely they are to retain the belief that there are supports and services available in the community that could assist them. This conclusion is entirely consistent with the experience of Wintringham, which has found that it is invariably the most recently homeless of our clients who retain a positive outlook and optimistic view of the future.

A recommendation arising from this research is the clear and undeniable need for an increase in the provision of affordable housing for the aged homeless and for those at risk of becoming homeless. Most long-term housing and support services within the community are unable to provide an appropriate level of skills, expertise or funding to adequately meet the needs of older homeless people, particularly with co-existing complex mental health issues. The lesson from our practical experience and from the results of this study is that the provision of affordable housing and the existence of low and medium levels of specialised support can prevent vulnerable aged people from becoming homeless and the myriad compounding health and social issues that may occur as a result.

It is hoped that by gaining a better understanding of the causes of homelessness through this and subsequent studies, we can inform policy on developing strategies to selectively target programs to provide appropriate and timely support to older people who are homeless or at risk of homelessness.

The other clear outcome of this research is the need for specialised services to be developed or improved so that the aged homeless can gain access to them. In Australia, there is no clear department or policy area that has responsibility for the aged homeless. Elderly homeless people have a variety of needs, the responsibility for which falls across many government departments, including Housing, Health and Aged Care, as well as Social Security and Veterans’ Affairs for income and specialised support services.

There is a distinct lack of statutory provision for the older homeless population as well as a chronic shortage of higher-level supportive accommodation options, e.g. 24-hour staffed hostel or self-contained flat/independent living units with appropriate support. Future funding should be directed into strategies designed to address such issues as gambling, depression and substance abuse among elderly Australians. But most importantly, more could be done to support the development of proactive preventive programs such as assertive outreach and the identification of at-risk individuals at the point of entry into emergency medical services, the justice system and community services.


The partners in the three-country study were the Sheffield Institute for Studies on Ageing, University of Sheffield, and the Committee to End Elder Homelessness and the Elders Living at Home Program, Boston, Massachusetts, US. The Australian research was funded by the Commonwealth of Australia, represented by the Department of Family and Community Services contract for Research Services under the National Homeless Strategy: International Study of Older Homeless persons. The authors thank the homeless people and the staff of the organisations that participated in the research.