This paper focuses on the residential mobility of middle-aged persons, not yet retired, an understudied cohort in mobility research. From the 1950s to the 1980s, mobility studies pointed to mid-life as a settled stage in terms of family, work and housing. Recent demographic and social changes, however, have led to these years being typified by a wide gamut of living arrangements that have complicated decisions about, and patterns of, residential mobility. Using the life-course perspective, this paper suggests that the transition to ‘empty nester’ status will heighten mobility among this group of middle-aged persons relative to their counterparts in other living arrangements. The analysis uses a customised migration matrix from the Australian 2006 Census and identifies segments of 45–64 year olds most likely to have changed address since the previous census in 2001. CHAID statistical method partitioned the 45–64 year old population in Melbourne, Australia, into eight statistically significant segments based on life-course factors and mobility levels. Younger (45–54 years) mid-life empty-nesters changed residence at 1.4 times the mobility rate of all mid-life persons. For couples in this age group, empty nest status conferred a 13 percentage point ‘mobility premium’ compared with couples that still had children at home. The results contribute to a better understanding of housing consumption among mid-life households and broader debates on access to affordable housing and processes of urban growth. Copyright © 2009 John Wiley & Sons, Ltd.