Authors: John Creedy and Ivan Tuckwell
This paper reports a reweighting exercise for the New Zealand Household Economic Survey, which is the basis of the Treasury's microsimulation model, TaxMod. Comparisons of benefit expenditures in a variety of demographic groups, along with population data, reveal that TaxMod estimates differ substantially from totals based on administrative data, when the weights provided by Statistics New Zealand are used. After describing the method used to compute new weights, the calibration requirements are reported. These relate to the age structure of the population and the number of beneficiaries for Unemployment Benefit, Domestic Purposes Benefit, Invalid's and Sickness Benefits and Family Support and Tax Credits. The revised weights and expenditure estimates are reported and the resulting distribution of income examined. The new weights are found to produce much improved expenditure estimates, without distorting the resulting income distribution. The effects of reweighting are demonstrated using a simple policy simulation.