Review of 2021 Census Topics

HSRG members (*) and other related agencies:
Australian Institute of Health and Welfare *
Council to Homeless Persons *
Department of Family & Community Services (New South Wales Government) *
Department of Health and Human Services (Tasmanian Government)
Department of Social Services (Australian Government) *
Housing SA (Department for Communities and Social Inclusion, Government of South Australia) *
Micah Projects Ltd *
Mission Australia *
NT Shelter Inc. *
Prime Minister & Cabinet (Australian Government) – Indigenous Housing Unit
RMIT University (Unison Housing Research Laboratory) *
Sera's Women's Shelter Inc. *
Shelter Tasmania Inc. *
Shelter WA * – incorporating comments from Department of Communities (Government of Western Australia); Centre of Social Impact, University of Western Australia; and Anglicare WA.
Swinburne University of Technology (Department of Social Science) *
Uniting *
University of NSW, City Futures Research Centre
Victorian Homelessness Network *

Accurate homelessness estimates are essential for designing and funding policies and programs that impact people and families experiencing disadvantage across Australia.
HSRG identified six key issues with the content of Census forms for estimating homelessness, and proposes solutions for consideration and testing:
To more accurately count young people who are couch surfing, the Census needs to ask about the reasons people are not at home on Census night.
To fill the gap in data on family homelessness the Special Short Form and NPD Form need to be able to capture information about family relationships
Allow self-identification of non-binary sex/gender
The definition of ‘supported accommodation for the homeless’ needs to be clarified, to improve accuracy and efficiency of data collection.
Identification of veterans who are an emerging homeless population.
To enable counting of persons not in a dwelling one or five years ago, this potentially lends to homelessness analyses of entries and exits.
Census enumeration processes also needs to be improved for accurate data collection.
Issue 1 - Where does this person usually live?
This question is essential to estimate persons experiencing homelessness, including:
persons who are in improvised dwellings, tent or sleepers out (sleeping rough)
persons staying temporarily with others (couch surfing)
persons in other temporary lodging
However, the ABS recognises that this methodology results in an undercount, particularly of youth homelessness.
Young people who are couch surfing are likely to put their parent’s address as their usual address, even if family conflict prevents them from returning there. This has also been raised as an issue for women and children fleeing domestic violence.
The question also does not differentiate between homeless persons staying in temporary lodgings and ‘grey nomads’, as both cohorts generally are enumerated in hotels, motels and caravan parks; have no or little income; and have no usual address. In the 2016 Census the ABS attempted to differentiate by encouraging homeless persons in temporary lodgings to write “None - Crisis” rather than “None”. This however was not included as an instruction on the form, was not well publicised, and did not account for the different terms used for this type of homelessness in different states. In fact, only 63 persons were added to the homeless count by reporting “None-Crisis”.
Proposed change
Tick-box response options should be added to this question for people with no usual address, with supporting instructions. For example: “No fixed address”, and “Tourist or traveller with no usual address”.
A question should follow-up for ‘visitors’ who are not enumerated at their usual address to identify reasons why they were away from home on Census night. This would better enable identification of couch surfers, as well as other groups, such as FIFO workers and domestic tourists. This question should be multiple choice, including an ‘Other (please specify)’ category. For example:
What was the main reason you were away from home?
Travel or holiday
Work
Education or training
Accessing services (e.g. health services)
Special occasion or event (e.g. wedding, reunion, concert, sport)
Wanted to visit friends, family or partner
Conflict or family issues at home
Home not currently habitable (e.g. natural disaster, renovations)
Housing is unaffordable
Other (please specify)

The Census is not an appropriate forum to directly ask about family violence or abuse. This could cause distress and that there is no way to ensure the wellbeing of respondents. Therefore, a broader category, such as ‘conflict or family issues at home’, should be considered and tested.
This question assumes that the person completing the Census form is aware of their visitor’s reason for being away from home. This may not always be the case, such as, if a young person was staying with a friend’s family. Collaborating with schools and service providers to educate parents and young people about youth homelessness could improve accurate reporting in the Census.
Should ‘Conflict or family issues at home’ prove difficult to answer, ‘Other’ could be a suitable response supported by education and awareness through service providers of how to report.
The ‘reason for visiting’ question could be further enhanced by including a direct question about dwelling type to enable more accurate classification of homelessness groups.
Issue 2 - Family Homelessness
The current methodology for homelessness enumeration in the Census does not enable family homelessness to be measured outside severe overcrowding. Understanding family homelessness is essential for policy and program design. For example, the service and supports needs of lone young person who is homelessness will be very different to a young person in a family who is homeless (e.g. Youth homelessness is one of the NSW Premier’s Priorities.
A ‘Special Short Form’ was designed to enumerate individuals sleeping rough so did not include relationship questions. In 2011 the ABS attempted using the Household Form to enumerate families sleeping rough, and in 2016 the ABS attempted to collect data to link family members’ Special Short Forms. Data on families sleeping rough collected using these methods was not of sufficient quality for the ABS to publish.
People who are temporarily staying with friends or family are enumerated using the Household Form. The Household Form collects relationship to Person 1 on the form, and uses this to derive all relationships in the household. This means that multiple families within one household can only be identified if they are related. Families temporarily staying with family friends cannot be identified. This also means that the ABS is undercounting the number of families experiencing severe overcrowding.
People experiencing other types of homelessness complete the Census using the Individual Form, which does collect data on family relationships.
Proposed Change
The ABS designs a family version of the Special Short Form that includes key questions on family relationships and allows for several people to be enumerated on the one form. This would be more efficient than using the Household Form, or completing several individual Special Short Forms, especially if the Special Short Form is moved to an online application.
Similarly, the ABS develops a Family Non-Private Dwelling (NPD) Form to enumerate NPDs such as hotels, motels and crisis refuges. Although this would result in a larger form, it would reduce burden on the managers of these NPDs (SFOs and PICs), as they would only need to record forms completed by each family, not each person.
Questions on family relationships on the Household Form should be adjusted to include non-related families living in the one household. So that this does not introduce unnecessary respondent burden, this could be done through follow-up questions for any person not related to Person 1 on their Census form. For example:
 Are you related to anyone in this household on Census night? Yes/No
 Who is the first person listed on this form that you are related to? Person 2, Person 3, etc
 What is your relationship to that person?
This additional information would enable improved derivation of housing suitability and its application to overcrowding measures
Issue 3 – Is the person male or female?
The current question wording does not reflect as a minimum, the Gender Standard Classification in Standard for Sex and Gender Variables, 2016 (Cat No 1200.0.55.012) where ‘Other’ is a third category. Under the Australian Government Guidelines on the Recognition of Sex and Gender, the Commonwealth Government expects Departments and Agencies to be collecting this data. Furthermore, the Census question does not define whether the concept of collection is sex, gender or sexual orientation.
In its current form the Census does not allow for any analysis of LGBTIQ homelessness in spite of recent Australian research showing that LGBTIQ individuals are at risk of homelessness and experience poorer mental health outcomes than their contemporaries. The Final Report of the GALFA LGBTQ Homelessness Research Project (September 2017) relied on analysis of General Social Survey (2014) and ‘Journeys Home’ longitudinal survey, with qualitative interviews, to report on LGBTIQ homelessness. This report recommended that Australian homelessness and housing policies should include LGBTIQ people as vulnerable sub-groups that require specific attention; and that data collection that includes sex, sexual orientation and gender identity should be mandatory and linked with homeless service funding agreements.
Proposed change
The Census should capture ‘Other’ as a minimum, as there currently is no method of counting gender diversity in the Australian population.
The HSRG acknowledges that LGBTIQ homelessness is an ongoing area of data interest, with growing policy interest. Improving information gathering on genders, sex characteristics, and sexual orientations would assist 6
research in this area. HSRG will defer to submissions from LGBTIQ community organisations for their advice on appropriate question wording, and their assessments against the criteria.
Issue 4 - People in Supported Accommodation for the Homeless
In this operational group, the ABS includes people in NPDs identified as “hostels for the homeless, night shelters, and refuges” on the Census Address Register and in dwellings flagged as Supported Accommodation Assistance Program (SAAP) accommodation by Governments and sector providers.
However, with the end of SAAP, what the ABS considers as supported accommodation is unclear and does not account for the increasing diversity and flexibility of the housing arrangements provided through homelessness programs. For example, short-term head leasing arrangements are becoming more common.
Proposed Change
The ABS, in consultation with HSRG, conducts a systematic review of different types of housing arrangements. From this, a clearer definition should be formed. The ABS should then explore whether it is possible to collect the information required to derive support accommodation through the Household Form.
For example, if the definition is ‘short-term housing provided by Government or a Community Housing Provider’, it may be possible to derive this from the current tenure and landlord questions if a question about length or security of the tenure arrangement was added.
Any definitional change will then be considered to ensure correct addresses, on supported accommodation lists, are provided by Governments and sector providers. In 2016, lists were asked to differentiate between accommodation types: emergency/short-term, transitional/medium-term, and long-term accommodation. This information can be explored further if suitable for estimation purposes.
Issue 5 – Veterans homelessness
The Census form does not have a question or option to select whether the respondent is an Australian Defence Force veteran. The 2018 State of Homelessness In Australia's Cities report of consolidated registry weeks found that 6% of homeless respondents were classified as veterans, 61% were rough sleeping, 43% had a serious brain injury or head trauma, and a much higher proportion than for non-veterans had a permanent physical disability that limited their mobility.
AIHW introduced a new question in November 2016 into their Specialist Homelessness Services Collection (SHSC) identifying clients who are current or former members of the Australian Defence Force (ADF). Results will be available from December 2018.
The Department of Veteran’s Affairs recently engaged UNSW and Adelaide University to research homelessness in the veteran community. Its aim is to collect data about homelessness in the veteran community so the Government can then better integrate support services available to homeless veterans with those offered by mainstream specialist homelessness service providers. This dataset will complement the new AIHW data. However Census estimates would further assist policy makers by identifying the locations and characteristics of the homeless veteran population, and whether the service delivery response is adequate.
Proposed Change
The ABS introduces a question similar to AIHW, to identify if a person had served in the Australian Defence Force for persons 18 years and above.
Are you a current or former Australian Defence Force member?
 Yes
 No
 not stated
 inadequately described
Issue 6 – Usual residence one or five years ago
The Census asks “Where did the person usually live one year ago?” and “Where did the person usually live five year ago?” For persons who had no usual address during these times is advised to write the address at which they were then living. This is not an appropriate instruction for those who were not in a dwelling because of homelessness. Asking to choose a ‘No fixed address’ tick box would be a more suitable option. It would enable analyses of longer-term (chronic) rough sleeping, and enable longitudinal analyses of housing and homelessness pathways, but also analysis on homelessness exits for those who were enumerated in a dwelling but were previously without a dwelling.
According to General Social Survey 2014, of persons who had experienced homelessness in the last ten years, 15.6% of people had experienced homelessness 1-2 years ago and 30.6% had experienced it 5-10 years ago.
While the suggested question wording might apply to persons who were in improvised dwellings (eg. Remote construction workers) or in tents (eg. campers) due to non-homeless reasons one or five years ago, they are more likely to report their usual address. In any case the estimation methodology would have filtered them out in previous homelessness estimates.
Proposed Change
Add ‘No fixed address’ tickbox to the ‘Where did the person usually live one/five year(s) ago” question.

The Census remains the single most important data source to assess the scale of homelessness in Australia. More recent policy initiatives and social trends have heightened the national importance of accurate homelessness statistics.
In 2017, The Commonwealth Government announced an intention to reach a National Housing and Homelessness Agreement (NHHA) with the State and Territory Governments. One aim of the NHHA is increased transparency and accountability for delivery on outcomes by State and Territory governments. This increases the national need for accurate homelessness data. The NHHA will commence on 1 July 2018.

Social impact investment is increasingly being used to fund homelessness and housing programs. ABS Census homelessness estimates are being used to model and set market prices under these investment models.
The increase in homelessness between 2006, 2011 and 2016 demonstrates that homelessness is a growing social issue. Homelessness estimates provide an evidence base for advocacy and program funding. The estimates are particularly important for identifying and developing programs to meet the needs of new and existing cohorts experiencing homelessness, such as older women; and those living in extreme crowding as a result of unaffordable housing. In this way, the Census homelessness estimates impacts the quality of life of some of the most disadvantaged people in Australia.

The Census homelessness data is anticipated to provide greater research capacity once it is successfully linked to the Multi-Agency Data Integration Project (MADIP). The current lack of a longitudinal data source to examine pathways in and out of homelessness is seen as an obvious data gap, in which MADIP is seen as one potential solution to address this shortcoming.

The Census is the only collection enabling an ongoing measurement of total homelessness across Australia. Administrative data sources such as AIHW’s Specialist Homelessness Services Collection (SHSC) reports on clients seeking services, and tends to exclude certain types of homelessness such as severe overcrowding, and those who do not engage with homelessness services. There is likely to be systematic bias in this data based on the locations of services, the cohorts services are aimed at, and differing ability of different cohorts to navigate homelessness services.

ABS survey data is also not adequate. The proportion of people in Australia experiencing homelessness on any one night is relatively low (at 0.5%) and is heterogeneously distributed across geographies. As policies are generally aimed at certain cohorts, and programs are typically funded to be delivered within specific geographies this creates a requirement for small group and area data.
Some of the key sub-populations that the Census enables disaggregation of are:
 Youth (alone and as part of a family)
 Older women
 Single parent families
 Aboriginal and Torres Strait Islander people
 People from CALD backgrounds
 People with a disability
Other key sub-populations that the Census does not enable disaggregation of are:
 Out-of-home-care leavers
 People who have left domestic violence
 People with experience of mental health
 People who have had contact with the justice system
 People from the LGBTIQ community
 People who are veterans (whether deployed or not)

However, the Census is not able to collect data to differentiate between all these groups. Furthermore, the Census is not an appropriate forum to ask about traumatic experiences, such as domestic violence.

The proposed modifications will improve ways of enumerating and estimating the overall prevalence of Homelessness in Australia and particularly the homelessness prevalence for young people, those fleeing domestic violence and Aboriginal and Torres Strait Islander populations. Improved estimation will allow greater planning location and distribution of services and other policy reform strategies to address the needs of these and other homelessness groups.

Homelessness is not a question directly asked on the Census form. It can be a very sensitive question to ask directly and even then can be interpreted in many ways. Instead, estimates of the homeless population are derived from the Census using both the characteristics reported in the Census and assumptions about the way people may have responded to particular Census questions.

For 2021, the proposed HSRG modifications to existing Census content should be simple enough to be understood and collected. With electronic reporting expected to be the main (and preferred) enumeration method in 2021, there are opportunities to simplify reporting through personalising the form and question sequencing. This should enable relationship questions to be added particularly to the Special Short Form.

As homelessness is not directly asked as a question on the Census form, the form in general should not be unacceptable for respondents to answer.

As noted in the summary, the Census is not an appropriate forum to directly ask about family violence or abuse, as this may cause distress to the provider, so care is needed to appropriately word responses to the proposed ‘Reason for visiting’ question. The proposal has the response option ‘Conflict or family issues at home’ to identify homeless couchsurfers. There is a risk however that people might incorrectly answer this question as the visitor is reluctant to inform the person filling out the form of their family circumstances, yet this modification is still a substantial improvement on the 2016 Census form.

Regarding overcrowding, some homelessness agencies and service providers report that some people completing the Census form are reluctant to identify the number of persons living in their private or public dwelling for fear that they may be evicted for breaching tenancy or visa conditions.

It is recommended that a strategy needs to be developed to identify these overcrowded dwellings and reassure occupants that information provided through the Census is anonymous and will not result in negative consequences. Homelessness and other welfare agencies have an important educative role in advance of the Census to reassure people about the purpose and anonymity of Census data.

Asking questions on sexual orientation to estimate LGBTI homelessness is likely to be offensive to parts of the community. Question sequencing on an online form may limit this to some extent. HSRG will defer this assessment to the submissions by the LGBTIQ community.

The proposed HSRG modification of the usual address question and including the ‘Reason for visiting’ question, will add some burden to the enumeration, processing, coding of homelessness. However the improved homeless identification should reduce the downstream effort in analysis and estimation, and improve the quality of the estimates for several published homeless groups. A positive consequence of the added question is it enables more precise estimates on journeys to work and holidays.

Expanding the Special Short Form and the Personal Form to capture family relationship information will increase provider burden and increase processing costs. From a provider perspective the length of an expanded SSF/PF is comparable to the longer questionnaire that is successfully asked of individuals sleeping rough participating in ‘registry weeks’. The additional variables sought are identical to that already asked on the Household Form so should be re-useable.

There will be an ongoing need for accurate data on homelessness. As the population increases, it is likely that the numbers of homeless persons will also increase. This will require a greater policy response, further investment by government and a strengthening of the homelessness support sector. Having a sound evidence base on which to base funding levels and locations will be imperative in ensuring government spending is efficient and effective.

The Census estimates of homelessness continue to inform Commonwealth and State/Territory policies on homelessness. The National Housing and Homelessness Agreement from 1 July 2018 will allocate funds of $4.6 billion over 3 years, and will include $375 million of new homelessness funding in which States and Territories are expected to match. A time series will enable a national and local assessment of the effectiveness of work programs.

Homelessness estimates feature in the Report on Government Services (RoGS) as performance indicator “Proportion of Australians who are homeless”, which is defined as the 'Number of Australians who are homeless' divided by the 'Number of Australians'. The state level estimates are used to proportionally distribute Commonwealth funds to each State and Territory.

The homelessness estimates are also used extensively by governments and community for state and local planning and delivery of homelessness services; and used by researchers to measure social disadvantage and poverty particularly for key vulnerable populations.

The ABS reports on the previous experience of homelessness from its General Social Survey (GSS), National Aboriginal and Torres Strait Islander Survey (NATSISS) and Survey of Disability, Aging and Carers (SDAC). Key information released include frequency, duration, where stayed, type of assistance sought and whether needs were met in the most recent episode of homelessness. However it is not comparable to the Census homelessness estimates as the scope are those who are no longer homeless and are living in private dwellings. Outputs are only at State and Territory level.
The ABS also reports on where persons stayed when fleeing a current or previous partner because of domestic violence in its Personal Safety Survey (PSS). This survey is not designed to measure prevalence of homelessness.
External sources of homelessness information include:
Rough sleeper counts – several inner city councils (Sydney, Parramatta, Adelaide, and Melbourne) conduct their own counts but they have differing methodology to each other and Census. In particular they are conducted at different times of the year and different times of the day. Yet they are useful to confirm the movements within the Census homelessness estimates.
Registry weeks – this assessment activity triages chronic rough sleepers with high co-morbidity into permanent supported housing. Its primary focus is to interview rough sleepers to determine their vulnerability through a diagnostic tool (VI-SPDAT). Some registry weeks do attempt to enumerate the prevalence of rough sleepers within their jurisdiction (eg. Inner Sydney) but are not comparable to Census data.
AIHW’s Specialist Homelessness Services Collection (SHSC) is a national homelessness data source that produces statistics on SHS agencies that are government funded to deliver accommodation-related or personal services to people who are homeless or at risk of homelessness. Statistics on supported accommodation services at a point-in-time is very useful to validate “supported accommodation for the homeless” estimates in the Estimating homelessness (ABS cat no 2049.0) publication - see paragraph 88 in the Explanatory Notes. However SHSC is not entirely comparable as it does not include reporting from non-SHS funded agencies so cannot be used for direct substitution (apart from methodological and statistical differences eg. metadata definitions and availability, reporting practices of providers). Notwithstanding, the SHSC may have complementary qualities worth exploring further, and potential changes within the National Housing and Homelessness Agreement may increase its utility as a future administrative data source.
‘Journeys Home’ survey was funded by DSS and managed by Melbourne Institute. It was a longitudinal dataset of homelessness that discontinued in 2015. It aimed to identify the factors leading to homelessness and the support strategies required to exit from it, rather than measuring prevalence of homelessness.

As outlined above, there are sources of homelessness data but these are either incongruent or unsuitable as alternative sources to the Census in offering a complete measure of homelessness prevalence in Australia.

Other content-related issues raised by HSRG members include:
Include question on Special Short Forms about duration/length of homelessness. (Western Homelessness Network)
New question on second residences could help filter out travellers, commuter workers, grey nomads (2016 HSRG submission)
Direct question on dwelling structure/adequacy to ensure mail-out areas have correct information (2016 HSRG submission)
Other issues not related to content raised by HSRG members include:
Frame
Separate 'Improvised dwellings, tents and sleepers out' into components including cars/vans, makeshift shelters; and other dwellings not meeting building standards (2016 HSRG submission)
Propose changing wording of "Q57. Housing Cooperative; Community or Church group" to "Community Housing provider, Housing Cooperative, Church group and other social landlord or service provider" to reflect the range of housing providers (CHP). Moreover, community housing providers will become more active as the Commonwealth Government introduces reforms to private affordable housing. HSRG expects Housing submissions to advocate this content change.
Planning/Collection
Refinement of special Census enumeration strategies for accurate collection of homelessness data, ie. sector recruitment, training, field support/materials, engagement, promotion/education (Uniting, CHP, Western Homelessness Network)
ABS to work with homelessness sector to develop and implement a National Homelessness Enumeration Framework/Strategy and Implementation Plan for the 2021 Census, and conduct an Evaluation of Framework/Strategy and Plan post-Census (Shelter WA, CHP, Western Homelessness Network)
Acknowledge diversity among supported accommodation for the homeless. Need to differentiate length of tenancy to be consistent with ABS definition, as different policy responses needed for emergency /short-term accommodation and medium/long-term accommodation (ABS, Tas DHHS)
ABS to engage with other community-based organisations, beyond SHS providers, to broaden "on the ground" contacts to improve enumeration of rough sleepers eg. local governments, park rangers (Shelter WA)
ABS work with State and Territory local councils (especially inner city areas), and local service providers to identify registered and unregistered boarding houses, if lists are not publicly available (DSS, Shelter WA, Swinburne)
ABS work with State/Territory housing authorities and the homelessness sector to develop a clear and consistent approach to develop supported accommodation lists prior to 2021 Census (DSS, Shelter WA)
Achieve full representation of Aboriginal and Torres Strait Islander homelessness and reduce the incidence of 'not stated' response to Aboriginal and Torres Strait Islander origin question for the homeless population (UNSW City Futures)
To improve enumeration of persons in overcrowded dwellings, suggest no limits on the online Census form. Currently a maximum of 10 people allowable for online form and 6 persons on Census Household form, although can request additional forms for filling (Shelter WA)
Estimation
Revisit estimation of overcrowding to ensure adherence to element of 'lack of privacy and social space' (DSS, Swinburne)
Consider income and sources of income as alternate means for accommodation, to improve estimation of homelessness (2016 HSRG submission)
Other
 data linkage to health, justice and social service data (Mission Australia)
 Investigate utility of AIHW SHS point-in-time data, for supported accommodation and rough sleepers presenting at services (ABS)
These suggestions were not included in the main submission as they are not related to form content, and will be considered during the inclusive strategy development phase.