APPENDIX 2 UPDATING THE WEIGHTS AND STRATIFICATION VARIABLES IN THE RPPIS


INTRODUCTION

1 The purpose of this appendix is to describe the process used to update the weights and the variables used to stratify the indexes. This process ensures the RPPIs continue to (a) provide a robust and accurate macroeconomic indicator of residential property price inflation; and (b) support the compilation of the non-financial assets component of the Household Balance Sheet in the Australian System of National Accounts (ASNA).

2 Every five years, following the availability of data from the Census of Population and Housing, an index review is undertaken to update the quantities of dwellings that underpin the weights of the indexes. It also provides an opportunity to update the variables used to stratify the indexes. The most recent index review occurred in relation to data from the 2011 Census and has resulted in the introduction of Series 3 of the indexes from the December quarter 2013 issue.


STRATIFICATION

3 The RPPI, HPI and ADPI all use a stratification approach to control for changes in the composition of properties bought and sold over the reference period.

4 The method of stratification used for Series 3 has not changed as a result of the index review. This means that suburbs continue to be grouped into strata based on their long term median price (the median price of the suburb between the 2006 and 2011 Census) and their Socio-Economic Index for Areas (SEIFA) score (2011) on the Index of Relative Advantage and Disadvantage. This approach balances homogeneity of suburbs within the same stratum with sufficient sales observations to construct reliable measures of price movements. For further information on the stratification approach used in the Indexes, see Information Paper: Forthcoming Changes to House Price Indexes: Eight Capital Cities (cat.no.6416.0.55.002) and House Price Indexes: Concepts, Sources and Methods, Australia, 2009(cat. no. 6464.0).

5 As a result of the updated variables for stratification, the number of strata in each city has changed for both the HPI (except Darwin) and the ADPI.

Table A2.1 Number of strata, Series 3 (from the June quarter 2013)
	HPI	ADPI
Sydney	23	12
Melbourne	19	13
Brisbane	13	9
Adelaide	13	8
Perth	13	10
Hobart	8	4
Darwin	6	7
Canberra	10	8

6 For information on the number of strata for the previous series of indexes, please consult House Price Indexes: Concepts, Sources and Methods (cat.no.6464.0) and the feature article, Experimental Other Dwellings Price Index in the June quarter 2012 issue of House Price Indexes: Eight Capital Cities (cat. no. 6416.0).


NEW WEIGHTS

7 The weights underpinning the indexes are based on the value of dwellings (including land) in scope of the indexes. The weights are expressed in terms of stock values. While the underlying quantities (i.e. number of dwellings) are held constant from period to period, the stock values are updated each period to reflect the changes in prices. The strata for the HPI and ADPI are re-valued each quarter by applying a price relative (i.e. the current period median price of the stratum compared to the previous period median price of the same stratum) to the value of the dwelling stock for that stratum to produce a current period stratum value. The current period values of each stratum are then summed to derive the current value of the total dwelling stock in the capital city. Index numbers are subsequently derived from the total values.

8 Over time the number of dwellings in a city will change. To maintain the indexes relevance, it is necessary to update the quantities which underpin the dwelling stock weights. This reweight utilised dwelling counts from the 2011 Census of Population and Housing to derive values to replace those calculated with data from the 2006 Census. The quantities have been valued with the March quarter 2013 mean prices to produce new weights for the indexes. A more detailed explanation of how weights are derived can be found in Chapter 7 of House Price Indexes: Concepts, Sources and Methods, Australia (cat. no. 6464.0)

9 The indexes are available for each of the eight Capital cities. The Capital cities in Series 3 are defined by the Australian Statistical Geography Standard (ASGS) level of the Greater Capital City Statistical Areas (GCCSA) as used in the 2011 Census. The new geography standard introduces changes to the capital city boundaries. For more information see Australian Statistical Geography Standard (ASGS) (Vol 1, cat. no. 1270.0.55.001). The following tables show the weights for Series 3 of the indexes.

Table A2.2 Series 3 Weighting Pattern, (a), percentage contribution to eight capital cities
	RPPI	HPI	ADPI
Eight Capital Cities	100.0	100.0	100.0
Sydney	35.6	32.1	45.8
Melbourne	28.7	29.2	27.2
Brisbane	11.6	12.7	8.6
Adelaide	6.8	7.3	5.3
Perth	13.1	14.4	9.4
Hobart	1.0	1.1	0.6
Darwin	0.8	0.7	0.9
Canberra	2.3	2.4	2.2
(a) Components may not sum to 100.0 due to rounding

Table A2.3 Series 3 Weighting Pattern(a), percentage contribution to RPPI
	RPPI	HPI	ADPI
Eight Capital Cities	100.0	74.1	25.9
Sydney	100.0	66.7	33.3
Melbourne	100.0	75.4	24.6
Brisbane	100.0	80.9	19.1
Adelaide	100.0	79.8	20.2
Perth	100.0	81.5	18.5
Hobart	100.0	84.2	15.8
Darwin	100.0	69.1	30.9
Canberra	100.0	75.6	24.4
(a) Components may not sum to 100.0 due to rounding

10 For information on weights for the previous series, please see House Price Indexes: Concepts, Sources and Methods (cat.no.6464.0) and the feature article, Experimental Other Dwellings Price Index in the June quarter 2012 publication of House Price Indexes: Eight Capital Cities (cat. no. 6416.0).


IMPLEMENTING THE NEW WEIGHTS AND STRATA

11 Series 3 indexes commence in June quarter 2013, which is the most recent quarter of the final series. The final series consists of index numbers produced with only State/Territory Land Titles Office or Valuers' General Office in each capital city (VGs) data and are not subject to revision (see paragraphs 26 and 27 of the Explanatory Notes for further detail).

12 The new price index series with updated weights and variables used to stratify the indexes are joined to the old series to form a continuous series via a process known as chain linking. At the link period, which in this instance is the March quarter 2013, new dwelling stock weights and strata are introduced in parallel to the old basis and median prices are calculated using both the new and old strata. The published index number for the link quarter is produced on the old basis, however movements and index numbers from this quarter on are derived by moving forward the new link period values with price relatives of the new strata.


IMPACT ON REVISIONS

13 Estimates for the two most recent quarters are preliminary and subject to revision. In this issue, the revisions to the June and September quarters 2013 not only reflect changes in the composition of mortgage lenders and VGs data in the set of prices used to derive medians (the usual reason for revisions) they also reflect changes to the weights of the index, and changes arising from updating the variables used to stratify the indexes.


IMPACT ON UNSTRATIFIED MEDIANS AND NUMBER OF TRANSFERS

14 The number of transfers and the unstratified medians are not directly impacted by the process of updating the weights and the variables used to stratify the indexes. However, users should note the change in geography from the Australian Standard Geographical Classification to the Australian Statistical Geography Standard when interpreting these data and exercise caution when comparing data across time.