Alternative Options for Living Arrangement Models: A Sensitivity Analysis

Abstract

Projections of the living arrangements structure of the Swedish population are carried out using actual marital status (in contrast to legal marital status) as the main criterion. The analyses and projections are based on an excellent Swedish data set which gives unique information on consensual unions as compared to marriages. By comparing demographic differentials in fertility, mortality, and couple formation and dissolution, both by actual and legal marital status, it is concluded that de facto marital status is by far more discriminatory than de jure status. Fertility among cohabiting women is found to be higher than usually recognized. Several alternative status selections as input to the multi-state projection model are suggested and tested, ranging from the ‘minimum model’ which only considers actual marital status to the ‘complete model’ which considers all actual and legal marital status combinations. It is found that, for the projection of living arrangements, the inclusion of legal marital status information is superfluous. Based on conservative assumptions, the proportion of consensual unions among all unions is projected to increase from about 23 in 1985 to 34 per cent by 2020 among the working age population, and from three per cent in 1985 to 12 per cent by 2020 among the elderly population (aged 65 years and over). While the number of consensual unions will increase remarkably, even by 2020 the majority of Swedes would still live in marital unions. Projected developments in the legal marital status distribution in no way reflect expected changes in actual marital status, i.e. in living arrangements.