Imputing Rent for Owner Occupied Dwellings in Household Income Statistics
Imputed rent for owner occupied dwellings (OODs) is a significant component of the international standards for household income and expenditure statistics. The ABS has not previously released household level estimates for the imputed rent of OODs in its household income and expenditure statistics, although household sector estimates have been included in the Australian National Accounts for many years.
Conceptually, the inclusion of imputed rent treats owner occupiers as if they rent their home to themselves, thus simultaneously incurring rental income and expenditure. Analysts using this data in income distribution studies can gain additional insights into the economic wellbeing of the population. Its inclusion allows more direct comparison of income across different housing tenures and over different life cycle stages.
Analytical Services Branch have been working with Living Conditions Section on a project to impute rent for OODs in ABS household income statistics, using data from the Household Income and Expenditure Survey (HIES) 2003-04 and the Survey of Income and Housing (SIH) 2005-06.
As owners don't actually pay rent, the key issue is what data is available to enable imputation of a rental value for each OOD and what methodology should be applied to estimate it. HIES and SIH collect data on the rent paid by renters and the housing costs incurred by owner occupiers i.e repairs and maintenance, rates, body corporate fees, insurance and mortgage interest payments.
We have used the 'market value' approach recommended in the international standards for household income and expenditure statistics, International Conference of Labour Statisticians, 2003. It calculates a net imputed rent income by estimating an equivalent market rent and subtracting the expenses incurred in earning that income i.e any expenses that would be paid by a landlord in the case of a rented dwelling.
Weekly rent paid by renters is modelled against a range of household and location characteristics, including number of bedrooms, type of dwelling, income and area characteristics measured by state, SEIFA and median rent postcode deciles. To account for selection bias caused by fitting a model to renters, and applying it to owners, we have used the Heckman model recommended by the European Statistical Agency (Eurostat). The model is then fitted to OODs, predicting their market rents, given their particular household and location characteristics.
The project is currently producing provisional estimates and comparing the results obtained with the sectoral aggregates released in the national accounts. It will also be undertaking analyses of the impact of the inclusion of the OOD rent estimates on income distributions generally, and for specific population sub-groups, in order to assess their suitability for release.
For more information please contact Jonathon Khoo on (02) 6252 5506.