Using administrative data sources to estimate immigrant flows to small areas

Nigel Swier, Office for National Statistics, UK
Ruth Fulton, Office for National Statistics, UK
Ian McGregor, Office for National Statistics, UK
Sofie De Broe, Office for National Statistics, UK
Amanda Sharfman, Office for National Statistics, UK
Rebecca Tinsley, Office for National Statistics, UK
Nicola White, Office for National Statistics, UK
Simon Whitworth, Office for National Statistics, UK
Alistair Dent, Office for National Statistics, UK
Jennifer Ford-Evans, Office for National Statistics, UK
Helena Howarth, Office for National Statistics, UK
Ben Winkley, Office for National Statistics, UK
Peter Youens, Office for National Statistics, UK

National Statistics Offices are increasingly looking at the opportunities of using administrative data sources to supplement or replace surveys and censuses. The Migration Statistics Improvement Programme has been led by the Office for National Statistics in the UK. One of the aims of the Programme has been to improve the estimates of long-term international migration at a local level through the use of administrative data from employment, education and health administrative systems. International immigration figures for England and Wales are estimated using the International Passenger Survey (IPS) that asks immigrants for their reason for visit (‘work’, ‘study’ or ‘other’). This paper presents the method built on an approach proposed by the University of Leeds using a range of administrative sources. This is an approach that combines national level data from the IPS, which is specifically designed to meet the UN definition of a long-term migrant, with administrative data sources to provide estimates of migration that provide greater richness for small areas. Data linkage is used to minimise double counting across sources. This paper outlines the distributional model that has been developed and discusses how these methods could be used in the future to provide Census-type information on a regularly basis.