The Australian Bureau of Statistics (ABS) has for the first time released official estimates
of the prevalence of homelessness in Australia. The estimates, which are based on the ABS’s new definition of homelessness
and methodology for estimating homelessness using census data
, are provided for homelessness at the time of the 2001 and 2006 censuses. Estimates from the 2011 census are to be published on 12 November 2012.
The ABS found that, as at the 2006 census, 89 728 Australians were homeless. This represents 0.5 per cent of the Australian population at that time and a rate of 45 homeless people for every 10 000 persons. At the 2001 census, the ABS estimates that 95 314 people were homeless. As such, according to the ABS, between 2001 and 2006 there was a 6 per cent decrease in the number of homeless Australians, and the rate of homelessness fell from 51 homeless people per 10 000 to 45. The main factor contributing to this decline was a fall in the boarding house population.
The ABS’s estimates differ substantially from what were previously assumed to be the most accurate and reliable data on homelessness in Australia—those estimates produced by Chamberlain and McKenzie, and published in Counting the Homeless reports. According to Counting the Homeless data, in 2001
there were 99 000 homeless people whilst in 2006
there were 104 676 homeless. Hence, not only are the ABS homelessness estimates lower than those furnished by the Counting the Homeless figures, but they also contradict the finding that there was an increase in homelessness between 2001 and 2006 of 4.8 per cent. So, the question is, what are the changes that led to this result?
As noted in a previous FlagPost entry
, the prevalence of homelessness depends on how it is defined. A narrow and literal definition of homelessness as ‘rooflessness’ results in a lower prevalence of homelessness than does a definition that encompasses insecure or inadequate housing. Definitions of homelessness not only determine who is to be counted as homeless but also inform the way in which they are counted (that is, the methodology used to measure homelessness).
If quality, comparable homelessness statistics are to be gathered over time and across data sources, then this demands a ‘robust, defensible and evidence informed definition of homelessness’, and one that is developed for statistical purposes. The ABS has developed just such a definition through a top-down and bottom-up approach. Thus, it has not only considered national and international statistical definitions of homelessness, but also the actual living situations of homeless people (drawn from people who work in the homelessness sector) that could be captured within the definition.
According to the new ABS definition, a person is homeless when they do not have suitable accommodation alternatives and 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 does not include people who are ‘at risk of homelessness’ nor does it include housing situations that put people at risk of homelessness. The definition is broader than Chamberlain and McKenzie’s in that it seeks to incorporate the critical elements of ‘home’, rather than focusing on an historically contingent accommodation standard (‘a small rental flat—with a bedroom, living room, kitchen, bathroom and an element of security of tenure’). But at the same time it is more tightly focused and less ambiguous in its application through its concentration on a range of different actual living situations.
Following an extensive review of the methodology underpinning Counting the Homeless figures
, the ABS refined a new methodology that is intended to be consistent with its new definition of homelessness, align with Census questions and field procedures, and be consistent and repeatable across censuses. While the ABS’s official methodology builds on the work of Chamberlain and McKenzie and is quite close to it in places, there are some key differences between the two approaches.
Perhaps the main difference is that the ABS has removed imputed estimates of homelessness where there is insufficient evidence to support their inclusion. This applies particularly to people enumerated in improvised dwellings as either being at home or with no usual address (such as construction workers, road workers, owner builders, hobby farmers and new migrants). The ABS has also tightened the rules for the homeless grouping of ‘persons in temporary lodging’ in line with the new definition of homelessness—that is, to remove people who may have chosen to live under these circumstances or who have accommodation alternatives. The ABS has also included the new homeless group of ‘persons living in severely crowded dwellings’. Despite the ABS’s best efforts, we will never be able to definitively quantify the number of homeless Australians.
Homeless people are inherently difficult to count in a census which relies on dwelling-based enumeration. Some homeless people will not be enumerated in the census and there will be under- or over-estimation of homelessness as a result of there being insufficient information to be certain whether or not certain people are homeless. The ABS acknowledges, for example, that the official methodology is likely to underestimate the extent of homelessness among some groups (young people, Aboriginal and Torres Strait Islanders and people fleeing domestic or family violence). However, in the interests of transparency and consistency, and in the absence of a robust methodology for generating estimates for these populations, the ABS has focused on using what is the most appropriate methodology from a statistical perspective.
Thanks in large part to the ground-breaking work of Chamberlain and McKenzie, around which a ‘loose consensus’ had developed, Australia has for some time been a world leader where it comes to estimating the extent of homelessness. With the introduction of an official definition and methodology for counting the homeless, this situation is likely to be further improved. As noted above, we now have an official estimate of homelessness in Australia that will be refined and improved over time. With significant research
currently being conducted into the lives of 1500 Australians who are homeless or vulnerable to homelessness, we will not only have a better grasp of the extent of homelessness, but also an improved understanding of the nature and causes of homelessness.