Administrative Longitudinal Data: Architecture, Governance, and Harmonisation

This chapter delineates the design, governance, and use of longitudinal administrative datasets, positioning them as complements to survey panels. Using PLIDA, LAD, and LIAB as exemplars, we contrast architectures and consent regimes, highlighting implied-consent frameworks and Five Safes controls that enable secure reuse at the population scale. We define recurring methodological risks such as classification drift, linkage error, and coverage bias, and prescribe harmonisation workflows, calibrated weights, and variance estimation. To support cross-system comparability, we advocate metadata and dissemination. Cases demonstrate policy salience across income support, child benefits, and labour markets. This chapter offers a principled toolkit for ethical inference from administrative panels.