Small-Area Incomes: Their Spatial Variability and the Relative Efficacy of Proxy, Geodemographic, Imputed and Model-Based Estimates



Williamson, Paul ORCID: 0000-0002-8836-6570
(2016) Small-Area Incomes: Their Spatial Variability and the Relative Efficacy of Proxy, Geodemographic, Imputed and Model-Based Estimates. APPLIED SPATIAL ANALYSIS AND POLICY, 9 (4). 463 - 489.

[img] Text
ASAP Small Area Incomes revised.docx - Accepted Version

Download (698kB)

Abstract

This paper uses data from a UK Census Rehearsal to explore the problem of small-area income estimation. First, the nature of the problem is revisited through an examination of the way in which incomes vary spatially. Residential rather than labour market sorting is found to be the dominant driver; and the rich are found to exhibit greater spatial segregation than the poor. Even so, location is shown to capture only a small fraction of the overall variation in incomes. Second, the performance of competing small-area estimation strategies is assessed, uniquely comparing proxy, geodemographic, imputation and model-based estimates; and validating all of these against directly observed values. An area-level model, ecological regression, performs best. Unit-record imputation approaches capture similar levels of spatial variation in mean income, but have higher variances and greater systematic biases. The same can be said of a simple univariate proxy (% professionals), which even so proves surprisingly effective.

Item Type: Article
Uncontrolled Keywords: Income distribution, Income segregation, Imputation, Synthetic estimate, Ecological regression
Depositing User: Symplectic Admin
Date Deposited: 08 Mar 2017 16:00
Last Modified: 09 Jan 2021 07:27
DOI: 10.1007/s12061-015-9163-1
Related URLs:
URI: https://livrepository.liverpool.ac.uk/id/eprint/3006263