Top Incomes and Inequality in Australia: Reconciling Recent Estimates from Household Survey and Tax Return Data

We systematically compare differences in recent Australian income inequality estimates derived from tax records and survey data. We use customised tax tables provided by the Australian Taxation Office (ATO) to more precisely measure Australian top incomes as conceptualised in the international top incomes literature. We then document the inability of Household, Income and Labour Dynamics in Australia (HILDA) survey data, as well as the ATO 1% sample tax files, to replicate our top income share values. In contrast, accuracy is substantially improved using the ATO 2% sample introduced in 2011-12. Using HILDA data with enhanced income values from the 2% sample we better track top incomes. More importantly, we are able to use alternative income concepts and income-sharing units to provide more consistent cross-national comparisons than are currently found in the top incomes literature. Furthermore, we improve standard summary measure estimates of inequality found in the traditional survey-based Australian literature. Our findings show great potential value in linking tax records and survey data, but also show that this value is severely limited when using the 1% tax sample available for 2003-04 to 2010-11. Whereas the 2% sample substantially improves top incomes estimates, data without censored and perturbed income variables are needed and could be provided without unduly threatening tax filer confidentiality.