In this paper we explore the issue of skills mismatch for university graduates using time series data derived from the HILDA survey. We compare the incidence of overskilling amongst graduates and the general population in the context of developments in the labour market since 2001 and demonstrate the cyclical variation in overskilling. By considering the strength of mismatch as well as a separate identification of skills-job match, we identify their different behaviours over the business cycle and discuss some of the reasons for these variations. We also discuss the differences in overskilling between males and females and across broad age groups. Industry decomposition of overskilling allows also to identify major contributors to the overall skills-job mismatch. Finally, a simple model of the relationship between overskilling and the tightness of the labour market is presented and we explore its applicability to forecast the incidence of overskilling. We conclude with a discussion of how the knowledge about the behaviour of overskilling over the business cycle can be utilized in policy development and offer some suggestions for further work.