European Journal of Homelessness. Volume 14(3). pp. 63-85

Discussions of homelessness measurement methodologies have largely focused on the generation of primary data, for example point-in-time counts. Though there is long standing tradition in the use of administrative data for measuring homelessness, relatively little examination of administrative data as method exists, i.e. the set of socio-technical practices through which administrative data are generated. This paper undertakes an internationally informed review of 50 administrative data systems in order to deconstruct these systems and stage a methodological discussion. Uniquely, the review included systems from other policy fields outside of homelessness, including health and education, in order to learn from wider data practices. The discussion elaborates on six key design considerations driving administrative data systems, including; function; data architecture; data quality; ethico-legal considerations; privacy preservation; and data access and accessibility. To conclude, we outline what an ideal data system would look like in order to improve the potential use of administrative data to measure homelessness and our response to it, but, more importantly, in mobilising data more effectively in order to facilitate research and operational uses of data. The six design elements can inform future homelessness administrative data systems, whilst also sensitising researchers and users of current administrative data to its (socially) constructed nature.