Simulation and quality of a synthetic close-to-reality employer-employee population

Linked employer-employee data are highly valuable for labour-market research but among the most sensitive datasets to release, because firms and individuals are both identifiable. This paper simulates a synthetic, close-to-reality employer-employee population that reproduces the linked structure between firms and their employees. It then evaluates the quality of the synthetic data through benchmarking indicators that compare key distributions and relationships against the original. Published in the Journal of Applied Statistics, the work shows how complex relational microdata can be synthesised for safe analytical use.