There is an ever growing demand to make use of administrative data for statistical purposes. Not only do they provide frames and valuable auxiliary information for sample surveys and censuses; systems of inter-linked statistical registers (ie registers for statistical uses) have been developed on the basis of various available administrative data sources to produce a wide range of purely register-based statistics.
The major advantages associated with the statistical use of administrative registers include reduction of response burden, gains of long-term production cost efficiency, and potentials for detailed spatial-demographic and longitudinal statistics.
Meanwhile, there is a clear lack of statistical theories for inference based on administrative data, which has its own and many types of errors. Register-based statistics, when viewed in relation to the historical development of survey sampling theory, appear currently very much at a pre-Neyman (1934) stage in terms of their maturity. A general issue for methodology research is the conceptualisation and measurement of the statistical accuracy of register data, which allow for application of rigorous statistical concepts such as bias, efficiency, consistency, etc, as one is able to do in other branches of the statistical science. As such, this is one of the emerging research areas.