The socioeconomic stratification of shared physical custody in Sweden: What drives the divide?
22.09.2026 kell 16:00
TLÜ, Uus-Sadama 5, M-552
Ettekande teeb: Jani Turunen
Associate Professor of Demography, Stockholm University
Associate Professor and Senior Lecturer in Social Work, Södertörn University
The socioeconomic stratification of shared physical custody in Sweden: What drives the divide?
Shared physical custody (SPC) has become an increasingly common post-separation family arrangement in many European countries. However, previous research consistently shows that SPC is socially stratified, with higher-income and better-educated parents being substantially more likely to share physical custody than disadvantaged parents. Different dimensions of socioeconomic status (SES) may influence the likelihood of SPC through distinct mechanisms. Higher income may facilitate the financial and logistical demands of maintaining two suitable homes for children, education may increase awareness of the benefits of continued parental involvement and foster more egalitarian parenting ideals, while occupational class may reflect differences in job flexibility that affect parents’ ability to accommodate alternating care schedules. Although education, income, and occupational class are all associated with SPC, these dimensions are closely interrelated, making it difficult to assess their independent importance.
This study examines the relative contribution of different SES dimensions to the likelihood of shared physical custody among separated parents in Sweden. Data come from the 2021 Swedish Generations and Gender Survey (GGS), linked to register information on income. The analytical sample consists of separated parents with at least one child under age 18. Associations between SES and shared physical custody are estimated using linear probability models, followed by a Shapley decomposition of the model’s explained variance (R²). This approach provides a fair allocation of explanatory power among correlated predictors by averaging their marginal contributions across all possible model specifications.