The study utilised micro-level longitudinal data from Journeys Home (JH) and housing market data from the 2011 Census to econometrically model the probability of being homeless as well as the probability of entry and exit from homelessness.

It found that men are prone to higher rates of homelessness both because they are more likely to fall into homelessness, as well as less likely to escape homelessness.

Long-term ill-health also predisposes people to homelessness, but diagnosis of mental health reduces likelihood of homelessness, possibly because diagnosis means people are more likely to receive assistance relative to those undiagnosed.

As expected, risky behaviour (drinking, smoking and drug use) raises the chances of entering homelessness, as does a previous spells of homelessness. Such people are at higher risk of becoming and remaining homeless irrespective of the condition of local housing and labour markets.

However, for individuals without behavioural issues, the risk of becoming and remaining homeless is more closely tied to the condition of local housing and labour markets. The chance of becoming homeless is greater in regions with higher median rents and slack labour markets.

Although individuals married or in a de facto relationship are less likely to enter homelessness, if they do become homeless there is a significantly lower likelihood of escape as compared to singles.

Exiting homelessness is much more likely for some groups than others, with younger people exiting more easily than older people. Current employment status does seem to be related to exits with some connection to the labour market better than none.

Optimal policy responses to homelessness need to address risky behaviours of some individuals (especially those with problematic alcohol and drug use). However, those without problematic behaviours might benefit from closer location to more affordable housing and job markets.