Reweighting of SIHC for use in MITTS

The aim of this project is to improve the starting point of simulation in MITTS-A, by making the simulated aggregate expenditure closer to the observed aggregate expenditure.

Although the Survey of Income and Housing Cost (SIHC) is weighted to represent the general population, the number of income support recipients was not taken into account in the weighting procedure. Given the focus on income support in MITTS, it is important to get the simulated number in the base data closer to the actual number. A way of improving the starting point is by reweighting the data so it adds up to the desired population level for the different payments. The under-representation of subgroups, which can be corrected by reweighting the sample, is only one of the possible sources of error. It should be noted that other sources of error, such as a serious take-up problem with benefits, cannot be fixed by this approach.

In addition to improving the starting point for years for which we have data, this approach could also be used for small corrections to update to the demographic characteristics of the population in a future year. For example, the latest data available represents the Australian population in 1997/98, whereas policy makers are likely to be interested in the effect of changes on the population in 2003/2004. The 1997/98 data could be reweighted so that the distribution of characteristics and the amount of expenditure on payments are similar to those in 2003/2004 and the weighted total number of persons in the sample adds up to the slightly increased population levels of 2003/2004.

More Information about Reweighting of SIHC for use in MITTS (pdf 138 KB)

The Melbourne Institute contact for this project is Dr Guyonne Kalb.

This project has been finalised in the third quarter of 2004 and the final report has been submitted.

The research paper arising from this project is:

No.	Melbourne Institute Working Paper Title
26/2004	Accounting for Population Ageing in Tax Microsimulation Modelling by Survey Reweighting, By Lixin Cai, John Creedy and Guyonne Kalb, Abstract PDF format (521K)