The Challenges of Combining Two Databases in Small-Area Estimation: an Example Using Spatial Microsimulation of Child Poverty

Spatial microsimulation techniques have become an increasingly popular way of fulfilling the need for generating small-area data estimates. However, the technique also poses numerous methodological challenges, including the utilisation of two different databases simultaneously to produce estimates of population characteristics at the local level. An important but neglected question is whether different distributions of key variables within these two databases may affect the validity of the spatial estimation results. This study uses the significant policy issue of small-area estimates of child poverty rates in Australia to examine this question. The different income distributions for families with children in the two databases and the consequent effect on child-poverty estimates are assessed, while the apparent validity of these synthetic small-area poverty rates is gauged