Assessing the agreement between microsimulated and register-based kin counts in Sweden

Estimating kinship networks is a data-intensive undertaking, usually conducted using empirical sources or demographic models. Although empirical data, such as population registers, provide observed individual-level information, they are limited by scarcity, truncation and survivorship bias. Demographic models, including microsimulation, require less detailed data, but they often minimally account for population heterogeneity, family similarity and multipartner fertility. This study examines the agreement of kin counts (from grandparents to grandchildren) between SOCSIM microsimulation outputs and Swedish population registers for cohorts born between 1915 and 2017. The results show that microsimulation closely approximates mean kin numbers and reasonably reflects parity distributions. Microsimulation underestimates kin for recent cohorts not affected by truncation, but provides a more accurate representation of kin for older cohorts with missing parent–child links. These findings highlight the utility of microsimulation for reconstructing kinship when only population-level rates are available, supporting its application in kinship analyses, both historical and projected.