Successfully addressing social inequalities requires moving from one-dimensional to multidimensional poverty measures, but evidence on Australia is still largely reliant on the former. Using panel data and counterfactual simulations, we examine the relative roles of material resources, employment, education, health, social support, community participation, and safety perceptions in explaining changes in multidimensional poverty in Australia between 2001 and 2012. We find that year-on-year absolute changes in multidimensional poverty are mainly driven by fluctuations in social support, community participation, and health. Social support, health and material resources increased relative poverty, whereas personal safety, employment, community participation and education reduced it. Changes in socio-economic returns to parental characteristics had also some impacts on changes in poverty rates. These findings constitute evidence that integrating non-income indicators of wellbeing in Australian policies aimed at addressing poverty would enhance their effectiveness and efficiency.