In recent decades, research has revealed the dynamic and often episodic nature of housing and homelessness among disadvantaged populations. Retrospective and multi wave longitudinal studies have quantified the number and duration of homelessness episodes experienced by individuals and families and identified the individual and societal level predictors of entry to and exit from homelessness. This has improved our understanding of how and why homelessness is experienced in different frequencies and intensities among different groups. Relatively less attention though has been placed on the types of housing support people utilise to exit homelessness, an important subject given the changing and often diminishing role of governments across many countries in the provision of public and affordable housing. This study utilises longitudinal data on a highly disadvantaged population in Australia to estimate probabilities of transitioning from homelessness to public and community (or social) housing, private sector housing and housing provided by family and friends and vice versa. Multistate life tables are used to estimate lifetime homelessness durations for each of these three housing states. The results indicate that homeless exit pathways are more volatile for people unable to access public housing, faced with higher probabilities of repeat homelessness episodes and longer lifetime homelessness. These findings point to a trade-off from reduced public housing investment between increased cost-savings to government and increased incidence and duration of homelessness and reliance on welfare services.