Housing is a complex, emotive and highly contested policy area. In recent decades, there have been significant conflicts over the policy directions pursued by national governments. In the vacuum created by this lack of consensus, policy-making has become circumscribed, reactive and highly volatile. As a result, governments have failed to provide adequate responses to what many perceive to be a worsening housing situation. This paper reviews the applicability of John Kingdon's Garbage Can model of agenda setting and alternative specification for understanding the complexities of policy-making in the housing context. Garbage can theories reject conventional 'policy cycle' models which envisage policy development processes as rational and underpinned by the logic of problem solving. For Kingdon, policy making is chaotic, random and frequently arational. There is a loose relationship between problems and the policy solutions offered by national governments. Using housing case studies from the Australian context, this paper explains the principles of the garbage can theory and evaluates their usefulness. It concludes that certain modifications are required to apply the model outside of the American institutional context. With certain adaptations, this paper argues that the garbage can model is a powerful descriptive and explanatory model for understanding complex housing policy-making processes. By providing a model for understanding the forces which drive irrational and reactive policy change, it is hoped that the groundwork will be laid for pursuing a more genuinely problem-solving approach in housing policy. The paper pays particular attention to current housing reform directions in Australia, including initiatives to marketise the delivery of housing assistance.