Unregistered Emigration and Over-Coverage in Norway's Population Register: A Capture-Recapture Hidden Markov Model Approach

Register-based population statistics increasingly replace censuses across Europe, but their accuracy depends on administrative activity reliably signalling residence. Over-coverage (absent individuals recorded as resident) is difficult to detect because registration is incentivized while formal departure is not, and it concentrates among mobile groups, such as migrants. We apply a capture-recapture hidden Markov model to estimate population size, over-coverage, and migration dynamics among foreign-born adults who entered Norway 2007-21. Using administrative data, the model jointly accounts for migration, death, residents generating no register activity, and absent individuals appearing in household-level records. Norway offers a rare validation opportunity: the Norwegian Tax Administration removes individuals deemed to have left without reporting departure. The model independently recovers 99.5% of these administrative emigrations, with inferred departure timing closely tracking administrative decisions. Over-coverage rises to approximately 4% by 2021, substantially below a comparable Swedish application; without Norway's administrative correction, it would exceed Swedish levels.