This project involved descriptive analysis and modelling techniques to understand the structural drivers of homelessness, using data from three Census periods (2001, 2006 and 2011). The project utilised data from the ABS Census of Population and Housing, the Specialist Homelessness Service (SHS) Collection from the Australian Institute of Health and Welfare, and other relevant data.
Homelessness is found to be highly spatially concentrated, with 42 per cent of the nation’s homeless population concentrated in only 33 of the 328 regions (i.e. 10% of all regions) in 2011. By contrast, such regions received only 34 per cent of homelessness resources (in terms of SHS capacity). While this spatial mismatch of resourcing has lessened over time, redressing this imbalance is an important area for policy attention.
Homelessness is concentrated in particular urban areas and remote regions of Queensland, Western Australia, and the Northern Territory, with regions with higher shares of males, sole parents and Indigenous persons having higher rates of homelessness. There was weaker evidence of homelessness being linked to younger populations with this mainly being an urban phenomenon. Policy-makers might seek to target these demographic groups and regions.
Contrary to expectations, modelling results suggest that areas with high homelessness do not have a shortage of affordable housing (rather they tend to have an abundant supply). Furthermore, weak labour markets (i.e. with higher unemployment) were associated with lower per capita rates of homelessness. Tentative explanations for these puzzling results point to the fact that low-income, low-rent areas have a higher population at risk of homelessness. Relatively fewer people are at risk of homelessness in higher income, lower unemployment regions but they are more vulnerable if they exit housing (since there is a lack of affordable housing in those locations). Homeless people might also migrate to such areas seeking work. Provision of affordable housing in such suburbs might partially address this problem.