7.1 As part of the 16th series review the ABS has assessed the expansion of the geographical coverage of the CPI beyond the capital cities, and production of spatial price measures to complement the temporal CPI.
7.2 The Australian CPI is compiled and published separately for each capital city, with the national CPI compiled as the weighted average of eight capital cities. Prices are collected in capital cities only. These capital cities constitute the CPI reference population and represent 64% of Australian private households.
7.3 The HES sample covers expenditures for the whole of Australia, excluding very remote areas accounting for between one and two percent of the population. However, in constructing CPI weights, only expenditure data on the same defined scope of the CPI are used - i.e. households in capital cities only. The same geographical restriction applies to the collection of prices for the CPI.
7.4 Ideally the CPI population group should encompass all Australian households, but in practice this has not been feasible due to the substantial additional resources that would be required to collect prices outside the capital cities and compile reliable regional weights.
7.5 In past reviews the ABS has decided against extending the geographical coverage of the CPI on the grounds that the benefits were not seen as being commensurate with the costs. Empirical evidence from previous CPI reviews suggests that price movements in regional areas are not significantly different from those in metropolitan areas.
7.6 During the consultation phase of the review, user comments were invited on the need for regional indexes - expanding geographical coverage beyond capital cities, and their potential use.
7.7 Both in the submissions received and in the public seminars there was a perception that an expansion of the CPI coverage would automatically lead to a more representative measure of inflation. A number of submissions to the review supported the expansion of the geographical coverage of the CPI, generally suggesting a need to measure the living cost pressures faced by Australians in regional, rural and remote areas. Some submissions expressed the view that regional prices are often quite different from those prevailing in the capital cities.
7.8 While most AG members favoured a CPI with the widest possible geographical coverage, it was acknowledged that the costs of undertaking such an expansion would need to be weighed up with an assessment of the benefits. It was widely agreed that the ABS explore cost effective options to increase the CPI coverage.
7.9 While it is perceived that expanding the geographic coverage of the CPI would result in a more representative measure, there would be many conceptual and practical problems to overcome to ensure that the quality of the CPI is maintained. Some of these include: establishing regional price samples; obtaining robust weighting data; and ensuring a consistent approach to price collection.
7.10 The HES expenditure weights at state, capital city, and 'remainder of state' levels are sufficiently robust for CPI weighting purposes. ABS analysis using 2003-04 HES data suggests that at the national aggregate level, the expenditure patterns for areas beyond capital cities are not significantly different to capital cities. There would need to be quite a significant price movement differential between capital cities and regional Australia for there to be any impact on the headline CPI.
7.11 However, it is difficult to establish that price movements outside capital cities are different to those of capital cities. Previous ABS empirical analyses indicate that price movements (for food and fuel) outside capital cities exhibit similar price movements to capital cities. The ABS is undertaking some work with scanner data and is confident that, in time, this type of data will assist the ABS to assess price movement comparisons between capital cities and areas outside capital cities.
7.12 To address the perception that the CPI coverage is 'unrepresentative' the ABS will investigate cost effective options of producing state and territory based indexes for possible future use. Capital city indexes would be absorbed into the state and territory indexes, aggregating to a national index number. Initial investigations indicate that adopting a state-based approach would achieve a CPI geographic scope (i.e. population coverage) of around 98%.
7.13 Some users have requested specific regional indexes. The ABS does not see merit in constructing regional CPIs as obtaining accurate expenditure weights and price observations for specific localities outside capital cities would require a substantial increase to the HES and CPI price samples.
7.14 The ABS will explore cost effective options to expand CPI coverage beyond capital cities. However, this must be considered in the context of competing priorities within the ABS work program.
7.15 A spatial price index (SPI) enables price levels to be compared between geographic regions at a single point in time. Experimental work on SPIs by the ABS in 2005 showed the potential to produce estimates of price level differences across Australian capital cities. The research concluded that there were many conceptual and practical difficulties in constructing such measures and the results of these experimental SPIs were not officially released. Due to budget constraints development work on SPIs was cut from the Prices program in 2008.
7.16 During the consultation phase of the review, user comments were invited on the importance of spatial measures and their potential use.
7.17 The review identified moderate support for the ABS to make available valid spatial comparisons of price levels across localities. Users were of the view that these types of spatial measures would provide a valuable instrument for policy making in relation to funding decisions where regional differences in the cost of living are important.
7.18 There were a variety of requests for spatial data and these covered different geographies, population segmentations, and purposes. Most interested parties expressed a desire to use spatial measures for living cost comparisons between the capital cities, although there was also some demand for spatial measures for specific regions outside the capital cities.
7.19 The experimental work undertaken by the ABS on SPIs identified some conceptual and practical difficulties with using CPI price data to construct these measures. A key problem in spatial comparisons is the difficulty to precisely match products across regions using CPI data. Although items priced in the CPI basket are broadly similar, they differ in product specifications between capital cities to take account of local conditions. Only around 40% of the prices used in the CPI satisfied the across-city comparability requirement and were able to be used in the experimental SPI calculations.
7.20 In constructing the experimental SPIs the ABS acknowledged that an expenditure-based (fixed-basket) spatial measure provided an incomplete means of comparing living standards, however it represented the broadest measure able to be constructed at the time by the ABS.
7.21 If the ABS was to reinstate the SPI program the ABS would firstly need to develop a better understanding of the conceptual and methodological issues underpinning the construction of SPIs. Extensive consultation with users would also be required.
7.22 The ABS would, in constructing SPIs, need to evaluate any synergies with its commitment to participating in the Organisation for Economic Co-operation and Development (OECD) Purchasing Power Parities (PPP) program.
7.23 The ABS will undertake the development and publication of annual spatial price indexes (SPIs) for capital cities, subject to appropriate funding and consideration of competing priorities within the ABS work program.
7.24 The ABS will investigate annual SPIs with an expanded scope to include rest of state components, subject to appropriate funding and consideration of competing priorities within the ABS work program.