ESTIMATING HOMELESSNESS FROM THE CENSUS OF POPULATION AND HOUSING

The Census aims to count all persons in Australia on Census night (with the exception of foreign diplomats and their families). To maximise the quality of count of the Australian population, the ABS has a special strategy to enumerate some homeless populations that are hard to enumerate through the standard Census procedures. The Census is the only collection that goes to all persons in Australia, and is therefore the best source to get an estimate of the number of homeless people at any one point in time. However, ‘homelessness' itself is not a characteristic that is directly measured in the Census. Instead, estimates of the homeless population may be derived from the Census using analytical techniques, based on both the characteristics observed in the Census and assumptions about the way people may respond to Census questions.

With advice from the ABS Homelessness Statistics Reference Group (HSRG), the ABS has recently developed and published its statistical definition of homelessness. The methodology described in this paper articulates how the definition is operationalised using Census data to provide a point in time estimate or 'prevalence estimate' of homelessness. The statistical definition is the basis for making some of the judgements about the interpretation of people's responses to the Census questions to determine whether they were likely to be homeless on Census night.

In brief, under the ABS statistical definition of homelessness, when a person does not have accommodation alternatives they are considered homeless if their current living arrangement:

• is in a dwelling that is inadequate; or
• has no tenure, or if their initial tenure is short and not extendable; or
• does not allow them to have control of, and access to space for social relations.

The definition is described in more detail in the following Chapter.

There is no single variable in the Census which can accurately inform on homelessness. Data item categories such as 'improvised home, tents, sleepers out' and 'no usual address' will include both people who were likely to have been homeless on Census Night and people who were unlikely to have been homeless. For example, people travelling away from home for considerable periods and having no place in which they are likely to stay for six months of more in the year of the Census should correctly report 'no usual address'. People living in a shed as they build their new home will be enumerated as living in 'improvised home, tents, sleepers'. Such variables need to be considered along with other data collected in the Census in determining whether or not a person was likely to have been homeless on Census Night.

Commencing in late 2009 the ABS undertook a review of the methodology used by Professors Chamberlain and MacKenzie to compile their estimates of the homeless population, as published in the ABS Australian Census Analytic Program as Counting the Homeless, 2006 (cat. no. 2050.0). The work by Chamberlain and MacKenzie was ground breaking, and attempted to operationalise their 'Cultural definition of homelessness' using the Census of Population and Housing. This work has been important in informing the ABS approach to operationalising an official definition of homelessness to estimate those people enumerated in the Census that were likely to have been homeless on Census night.

ABS has developed six Homeless Operational Groups in assessing whether people enumerated in the Census were likely to have been homeless on Census night. These groups are:

• Persons living in improvised dwellings, tents, sleepers out
• Persons in supported accommodation for the homeless
• Persons staying temporarily with other households
• Persons living in boarding houses
• Persons in other temporary lodging
• Persons living in 'severely' crowded dwellings.

The first four of the categories are similar to the four groups used by Chamberlain and MacKenzie in their estimates of homelessness developed using Census data.

Estimation for the first four ABS homeless operational groups starts with the concepts and assumptions underlying the Chamberlain and MacKenzie homeless operational groups. ABS tested these assumptions and refined its own estimates to ensure that:

• they reflected the intent and design of the Census variables; and
• a homelessness interpretation of the Census data stands up to scrutiny against the concept and elements in the ABS statistical definition of homelessness (e.g. assessing the extent to which the concept of accommodation alternatives could be operationalised in the Census).

The fifth ABS group 'Persons in other temporary lodging' consists of persons who were classified by Chamberlain and MacKenzie as 'persons in boarding houses'. However these persons, who reported 'no usual address', were enumerated in a 'non-private dwelling' which was classified by its owner or manager as a 'hotel, motel, and bed and breakfast'. These properties were unlikely to have been boarding houses and are included instead in the new ABS group.

People living in severe overcrowding are considered to be in the sixth ABS homeless group because they lack of 'control of and access to space for social relations' (one of the key elements of the ABS definition of homelessness) and are considered not to have accommodation alternatives when remaining in such extreme living arrangements. Severe crowding conditions are operationalised in the Census as living in a dwelling which requires 4 or more extra bedrooms to accommodate the people who usually live there, as defined by the Canadian National Occupancy Standard (CNOS, See Glossary).

The six ABS homeless operational groups compiled from Census data are described in more detail in Chapter 5 Methodology. That chapter also outlines how each group aligns with the ABS statistical definition and what rules are used to operationalise each group.

Estimates of homelessness are important for providing a prevalence of homelessness on Census night and the characteristics of those who were likely to have been homeless. However, there are people whose living arrangements are close to the statistical boundary of homelessness, and estimates of such people who may be at risk of homelessness can be used to assist policy and service delivery to prevent people becoming homeless. Therefore, estimates, compiled from Census data for specific key groups of people who may be marginally housed, but who are not classified as homeless, will be presented alongside estimates of the homeless operational groups in the publication Census of Population and Housing: Estimating Homelessness (cat. no. 2049). Those groups are:

• Persons living in other crowded dwellings;
• Persons in other improvised dwellings; and
• Persons who are marginally housed in caravan parks.

More information about each of these groups, including the rules used to estimate them, are provided in Chapter 5 Methodology.

Later chapters in this Information Paper will cover:

• the ABS statistical definition of homelessness that underpins the methodology (Chapter 3)
• the limitations and issues with using Census data to estimate homelessness (Chapter 4);
• the key differences in methodology employed for estimating homelessness using Census 2006 data and the methodology using Census 2001 data (Chapter 5);
• the process that was undertaken to develop the methodology for estimating homelessness from the Census (Chapter 6); and
• the future directions for estimating homelessness from the Census (Chapter 7).

Previous Page

Next Page