Home after widowhood: a longitudinal study of residential mobility and housing preferences following a partner’s death

Abstract

Within the housing literature, questions regarding the residential mobility and housing choice of people who have recently transitioned into widowhood remain unanswered – especially when it comes to the differences between widows and widowers. This article utilizes unique Danish administrative panel data to investigate these questions through a case observing and following all widows and widowers between 50 and 90 years of age over a 25-year period. The results show that while both widows and widowers are significantly more likely to move in the years following a partner’s death, but also that widows are more residentially mobile than widowers. Furthermore, the results show that the transition into widowhood increases widows’ likelihood to move closer to their adult children, while both widows and widowers are more likely to downsize following a move. These results indicate that the death of a partner affects the residential situation of men and women differently, even within a Danish context that strongly emphasizes gender equality.

Keywords:

• Residential mobility
• housing preference
• widowhood
• ageing

JEL:

• J14
• R21
• R23

Acknowledgement

I would like to thank Cecilie Dohlmann Weatherall, Ismir Mulalic, Bo Honoré and Donald R. Davis for very important comments and helpful suggestions to this article. Furthermore, I would like to thank the participants at the European Regional Science Association (ERSA) 2019 for their valuable comments and feedback.

Disclosure statement

Nothing to disclose.

Notes

1 Care homes and senior housing are identified following the approach proposed by Statistics Denmark, where a residence is registered as a care home or senior housing unit if more than 6 people living within the same stairwell are above the age of 80 and if less than 25% of all people living in the stairwell are below 60 years of age (Statistics Denmark 2014).

2 As some people could be in a relationship with several different individuals throughout the period of observation and thus be part of multiple households in said period, the number of unique individuals in the sample is not exactly double the number of unique households but instead slightly lower.

3 The number of families does not match 1 to 1 with the number of individuals who have lost a partner, as in some cases both the husband and the wife from the same family have been matched with people who have lost a partner.

4 The proportion of individuals who have lost a partner in the final sample after matching is higher than 50%, as the matching was done with the replacement in the couples group, meaning that some individuals whose partner had not died were matched with multiple surviving spouses.