APPENDIX 2 COMPARISON OF GROSS IMPUTED RENT FOR OWNER-OCCUPIERS BETWEEN SIH AND ASNA
INTRODUCTION
This publication contains experimental household level estimates of the gross imputed rent for owner-occupied dwellings in Australia, derived from data reported in the 2003-04 and 2005-06 Surveys of Income and Housing (SIH). This is the first time the ABS has produced household level estimates of the gross imputed rent for owner-occupied dwellings, although household sector estimates have been included in the Australian System of National Accounts (ASNA) for many years.
The SIH and ASNA estimates of gross imputed rent for owner-occupiers have been developed for different purposes, but apply similar concepts and methods. The SIH experimental estimates are derived from data collected from individual households, and may be used to aid analysis of the distribution of household income and to make more useful comparisons across population subgroups. The ASNA estimates are produced at an aggregate level for the household sector and are presented within a national accounting framework, allowing comparisons with the corporate and government sectors in Australia and to the rest of the world.
This appendix describes the major scope and methodological differences between the estimates produced from SIH and ASNA, and compares the aggregate results for the two reference periods.
CONCEPTS AND METHODS
The concepts and methods underpinning the estimation of gross imputed rent for owner-occupied dwellings in SIH and ASNA have much in common, with both sources imputing a market value for the housing services accruing to owner-occupiers from the occupation of their primary residence (in ASNA a value is also imputed for any additional residences such as holiday homes). Both estimate the market value of the housing services accruing from the rental equivalent.
In producing the SIH imputations, the market value of the rental equivalent has been estimated by regressing actual rents paid by private renters on the characteristics of their unfurnished rented dwellings e.g. location, dwelling structure and number of bedrooms. The estimated coefficients have then been applied to the characteristics of owned dwellings to produce predicted values of the rental equivalence of these dwellings.
In producing the ASNA imputations, a benchmark is established using the Census of Population and Housing, which gives the number of owner-occupied and rented dwellings, and information about rents paid for rented dwellings. The imputed rent for owner-occupied dwellings is calculated by multiplying average private rents for unfurnished dwellings reported in the Census in various strata (defined by major urban, other urban, rural; cross classified by dwelling structure and number of bedrooms) by the number of owner-occupied dwellings in the same categories. For intercensal and post-census periods, the estimates are interpolated / extrapolated using a range of indicator data. Details are available in Section 04.2 of Australian National Accounts: Concepts, Sources and Methods (cat. no. 5216.0).
DIFFERENCES IN SCOPE
There are two scope differences which impact on the comparability of the SIH and ASNA estimates of gross imputed rent for owner-occupiers.
Firstly, the ASNA estimates include the gross imputed rent accruing to owner-occupiers not only from occupation of their primary residence, but also from any additional residences such as holiday homes. The ASNA estimates also include a portion of unoccupied dwellings. The SIH estimates capture only the imputed rent accruing to owner-occupiers from occupation of their primary residence. However, this scope difference can be quantified and is very small.
Secondly, SIH excludes households in collection districts defined as very remote. For most states and territories the exclusion of people in very remote areas has only a minor impact on the aggregate estimates because they constitute a small proportion of the population. The SIH estimates for the Northern Territory will be understated as very remote households account for about 24% of its population, but the scale of this exclusion is not clear. Because the majority of the population living in very remote households in the Northern Territory are living in discrete Indigenous communities where home ownership rates are very low, imputed rent estimates for home owners are not likely to be significantly underestimated for the Northern Territory. The estimates in this paper also exclude rental subsidies that may apply in the discrete Indigenous communities.
DATA USED IN COMPARISON
SIH is conducted throughout the financial year and respondents who rent their dwellings are asked to report the amount of rent they pay at the time they are surveyed. Therefore, the rent paid by private renters and the gross imputed rent for owner-occupiers are assumed to be average values for the year.
Elsewhere in this publication, the SIH estimates of gross imputed rent for owner-occupied dwellings are expressed in terms of a mean weekly value per household. To facilitate comparison with the ASNA estimates in this appendix, they are presented as Australian level annual aggregates. Multiplying the mean annual value per household for all households by the estimated number of in scope Australian households (7,735,800 for 2003-04 and 7,926,200 for 2005-06) provides the aggregate annual Australian values.
The ASNA estimates of gross imputed rent for owner-occupiers used in this appendix are those shown in table 60 of Australian System of National Accounts, 2006-07 (cat. no. 5204.0). Separate estimates of the imputed rent of owner-occupiers are only produced annually and are obtained by assuming a constant rate of change in the proportion of dwellings that are owner-occupied between Census benchmarks. For the post-census period, the rate of change observed for the previous intercensal period is used.
COMPARISON OF ESTIMATES
Table A2.1 shows that, in both 2003-04 and 2005-06, the experimental SIH estimates of gross imputed rent at the Australian level were within 2% of the ASNA estimates. In 2003-04 they were 1.6% lower, but in 2005-06 they were 0.7% higher. Between the two reference periods, the SIH estimates grew by 15.8%, while the ASNA estimates grew by 13.2%.
A2.1 Comparison of SIH experimental estimates with ASNA estimates of gross imputed rent for owner-occupiers |
|  |
| 2003-04 | 2005-06 | Difference | |
| |
| ASNA estimate | $60 511m | $68 519m | 15.8% | |
| SIH experimental estimate | $59 561m | $68 971m | 13.2% | |
| Difference | -1.6% | 0.7% | . . | |
| |
| . . not applicable |
The proximity of the two estimates can largely be attributed to the broadly similar methodologies applied and to the generally accurate reporting of private rents in both data sources used.
However neither set of estimates takes into account any quality differences in the stock of rented and owned dwellings and therefore both estimates possibly underestimate the 'true value' of gross imputed rent for owner-occupiers. If this problem were to be overcome in the SIH estimates, dwelling values would have to be captured for rental properties. Survey respondents occupying rental properties are unlikely to be able to provide a reliable estimate of the value of the dwelling.
The two sets of estimates are compiled for quite different purposes. The ASNA estimates are compiled at the aggregate household sector level. For compilation of the ASNA regular annual and quarterly estimates are required. The SIH estimates are compiled at the unit record level to aid detailed analysis of the economic circumstances and income distribution of Australian households. Estimates are potentially required whenever a SIH is conducted and are of primary use as supplementary items on the SIH Confidentialised Unit Record File (CURF).
Both the household level and the sector level estimates are subject to various types of error. SIH data are subject to sampling and non-sampling error. Sources of non-sampling error include non-response, errors in reporting by respondents or recording of answers by interviewers, and errors in processing the data. Sampling error arises from sampling variability, and from the weighting and benchmarking processes which are designed to adjust the results from a sample survey to infer results for the total in scope population.
The ASNA estimates are benchmarked to the Census of Population and Housing and are interpolated and extrapolated over time using a range of indicator data. The strength of the ASNA methodology is the comprehensive Census benchmark, which is subject to non-sampling error, but not sampling error. The main weaknesses of the method are due to limitations with the indicators used to extrapolate the estimates for future periods and the long time period between the availability of data for successive Censuses (five years) when new benchmarks are established.
Print Page
Print All