How far do internal migrants really move? Demonstrating a new method for the estimation of intra-zonal distance

Although population censuses in the United Kingdom provide the most detailed data on internal migration, they do not reveal with any degree of precision how far people travel when they make a permanent change of usual address. This paper makes use of alternative data from a large consumer survey to estimate distances of migration within England in the mid-2000s. Following a review of previous approaches to migration distance measurement and modelling, this paper seeks to evaluate the accuracy of conventional distance estimation methods and provide the basis for a new regression-based method of estimating intra-zonal distance. The new approach, which can be applied whenever different aggregations of the flow matrix and spatial units are available, is shown to result in significantly improved goodness of fit statistics for doubly constrained spatial interaction models.