Combining Household Income and Expenditure Data in Policy Simulations

Abstract

Analysis of the distributional impact of fiscal policy proposals often requires information on household expenditures and incomes. It is unusual to have one data source with high quality information on both, and this problem is generally overcome with statistical matching of independent data sources. In this paper Grade Correspondence Analysis (GCA) is investigated as a tool to improve the matching process. An evaluation of alternative methods is conducted using datasets from the UK Family Expenditure Survey (FES), which is unusual in containing both income and expenditure at a detailed level of disaggregation. Imputed expenditures are compared with actual expenditures through the use of indirect tax simulations using the UK microsimulation model, POLIMOD. The most successful methods are then employed to enhance income data from the Family Resources Survey (FRS) and the synthetic dataset is used as a microsimulation model dataset.