Producing the admin-based household estimates by number and size for each Local Authority and in England and Wales

Abstract

Objective
Timely and accurate households estimates in England and Wales are a key user need, however between Censuses, surveys cannot produce the same granularity. Our ambition in 2025 is to produce the first admin-based household estimates (ABHEs) by number and size for each Local Authority (LA) and in England and Wales.

Method
To devise our methodology we conducted an international literature review, a quality options strategy, a methodology options paper and from these produced a detailed plan. We utilise linked person-level administrative data, and also draw strength from survey data.

The first step of our plan is to produce admin-based household population estimates, aligned with the admin-based population estimates, for which we explored seven methods. The second step is to create the admin-based number of households by LA, again considering seven potential methods. The third step is then to produce the admin-based household estimates by size focusing on two potential methods.

Results
Presently we are carrying out the research and if results are available we will share. We will discuss our considered future interpretation and success measurements. In additional to comparing our ABHEs to Census and official surveys we are researching other administrative sources to produce a proxy for the number of households. Some of our considered methods for producing the household population include using linked administrative data, rather than census data like in current official methods, enabling comparisons between the two.

This research faces multiple challenges including: identifying statistical methods and statistical quality metrics when using linked administrative data and administrative data coverage, quality and access. We will explain how we addressed these issues, highlight remaining challenges, and outline our plans to tackle them further.

Conclusion
Producing ABHEs presents challenges but also involves innovation and ambitious research. This feasibility research aims to produce a proof of concept to later further enhance to Official Statistics in Development if successful. This presentation will demonstrate how administrative data could fulfil a key user need.