Household changes and housing consumption at older ages in Scotland: a comparison of two decades
Fiori, F., Graham, E. & Feng, Z. (2015) AISP conference, Giornate di Studio sulla Popolazione 2015, University of Palermo, Italy, 4 - 6 February 2015 [SLS]
The past few decades have seen significant demographic, social and economic changes that have resulted in increased diversity across individual life courses and housing careers. The study of residential relocation and housing conditions at older ages – in particular following changes in individual situations and household composition – is thus an important focus for research. Litwak and Longino’s (1987) identified three typologies of moves for older people: amenity-related or retirement moves; disability or health-related moves; and moves to institutions. Studies from the UK have supported this conceptualisation but Scotland has never been the focus of their empirical investigations. The aim of our study is to address this research gap. We investigate housing transitions in later adulthood in Scotland, covering the period when older adults are entering retirement and going through important changes in their households’ composition. The study addresses two main research questions: What are the key determinants of older adults’ residential moves? What are the key determinants of housing adjustments (downsizing/upsizing) among older individuals who move? We use data from the Scottish Longitudinal Study (SLS) for two decades (1991-2001 and 2001-2011) and adopt a repeated cross-sectional design to examine the circumstances of older people at the beginning and end of each period. The sample for each decade comprises individuals aged 55 to 69 and living in private households at the start of the decade. Separately for each decade, we observe whether individuals had changed their address (based on postcode of residence) by the end of the period. We use logistic regression to assess the relationship between the likelihood of a residential move and socio-demographic and housing characteristics measured at the beginning of the period, and household changes and other life events occurring throughout the period. Then, and only for individuals who changed address between two consecutive censuses, we observe whether the move implied any adjustment of their housing size. A multinomial logistic regression is used to contrast downsizing and upsizing to the base category same size, and to assess the influence of individual and family conditions (and their change over time) on the likelihood and the direction of housing adjustments. Both for residential mobility and housing adjustments, the models are extended to include contextual variables at the local authority level within a multilevel framework.