6469.0 - Outcome of the 16th Series Australian Consumer Price Index Review, Dec 2010  
ARCHIVED ISSUE Released at 11:30 AM (CANBERRA TIME) 06/12/2010  First Issue
   Page tools: Print Print Page Print all pages in this productPrint All  
Contents >> Analytical series >> Introduction

INTRODUCTION

6.1 As part of the 16th series review the ABS has assessed the need for, and relevance of, various supplementary indexes.


Seasonally adjusted estimates and an All groups excluding food and energy index

Background

6.2 The CPI measures contemporary movements in prices and reflects real world volatility that may occur, which can make it difficult for economic analysts to distinguish the volatility of short-term fluctuations from the underlying inflationary trend. To aid this analysis, the ABS produces a number of exclusion-based analytical series, which exclude certain items from the CPI basket. These analytical series are presented in Tables 9 and 10 of Consumer Price Index, Australia (cat. no. 6401.0).

6.3 Seasonally adjusted data are sometimes preferred in the formulation of economic policy and for economic research because they eliminate the effects of regular periodic events. The magnitude of seasonal effects can often mask the short-term underlying movements of the series. The ABS does not seasonally adjust the All groups CPI, nor does it currently produce a seasonally adjusted version of the CPI.

6.4 Seasonally adjusted CPI series are produced by a number of international statistical agencies (in addition to the headline CPI), including Statistics Canada and the United States Bureau of Labor Statistics (BLS). The United Kingdom Office for National Statistics (ONS) produces a seasonally adjusted version of an underlying inflation measure, the RPIY (All items Retail Price Index (RPI) excluding mortgage interest payments and indirect taxes).

6.5 The ABS produces two statistical measures of underlying inflation - the trimmed mean and the weighted median - using seasonally adjusted data, under the methodology developed by the RBA. These measures, which abstract from short-term volatility, are considered useful in identifying underlying trends of price level change.

User views

6.6 During the consultation process of the review there was strong demand for a seasonally adjusted CPI to be published alongside the other analytical measures. Analysts of consumer price inflation in the Australian economy indicated that a seasonally adjusted CPI would be a useful analytical measure to understand underlying trends in prices.

6.7 In its submission to the 16th series CPI review the RBA recommended that the ABS publish three additional analytical price series: the trimmed mean based on city-level data; the weighted median based on city-level data; and a trimmed mean inflation measure calculated using the year-ended distribution of price changes.

6.8 The ABS also received numerous requests to produce and publish an All groups excluding food and energy index as part of the exclusion measures in the analytical series. Many countries, including the United States and Canada, publish CPI data excluding food and energy as part of their analytical series, in addition to the headline CPI. Users stated that provision of an Australian 'CPI ex food and energy' measure by the ABS would allow for better international comparisons of inflation measures.

Evaluation

6.9 A 2008 ABS study, supplemented by more recent work, examined seasonality in the CPI. This study, using data up to December quarter 2007, found 64 of the 90 CPI expenditure class series to be seasonal. Although many expenditure classes showed seasonal influence, the All groups index did not exhibit any stable seasonality.

6.10 Given the user demand for such measures, the ABS sees merit in producing and publishing seasonally adjusted CPI estimates as part of the suite of analytical series.

6.11 The unadjusted CPI will continue to be the official headline measure. One of the long-standing features of the Australian CPI is that it is not revised in the normal course of events. However any seasonally adjusted CPI series would be subject to revisions as the estimate of the seasonal pattern is refined and updated each period as more data becomes available.

6.12 The seasonal adjustment will be undertaken using standard methodological approaches as used with other ABS publications. The seasonal factors calculated in producing seasonally adjusted indexes for the seasonal component series will also be used in the production of the trimmed mean and weighted median measures of underlying inflation (currently, the RBA undertakes the seasonal analysis and provides the ABS with the relevant factors for each quarter). The ABS will calculate seasonal factors for the 90 expenditure classes as used in the current measures only (i.e. for the weighted average of the eight capital cities). Production of additional seasonally adjusted measures, such as individual capital city indexes, will be subject to appropriate funding.

6.13 It is generally ABS policy to publish trend series derived from seasonally adjusted series. The resulting availability of a wide-range of underlying trend measures (the trend estimate, the trimmed mean, the weighted median and the 'excluding' measures) raises the question as to what trend measure the ABS would publish and promote as the CPI trend estimate. Final decisions on the presentation of seasonally adjusted analytical series and trend estimates will be detailed in the information paper to be published about one month ahead of the introduction of the 16th series CPI in October 2011.

6.14 The production of a monthly CPI (see Chapter 3) is likely to result in seasonal influences becoming more crucial in interpreting price change. At least three years of data would be required before a monthly seasonally adjusted CPI could be produced. The ABS does not support the imputation of past quarterly data to allow historic monthly seasonally adjusted data to be created.

6.15 To assist users in making international comparisons and to provide an additional analytical measure of underlying inflation, the ABS will produce an All groups excluding food and energy index as part of the analytical series from the September quarter 2011. Details of this new measure will also be made available in the information paper prior to the introduction of the 16th series CPI.

Outcome

6.16 The ABS will continue to produce analytical measures of underlying inflation. The current seasonal adjustment methods will be replaced by standard ABS seasonal adjustment methods. The ABS will produce seasonally adjusted estimates of the CPI, and of significant seasonal components (using standard ABS seasonal adjustment methodologies), from the commencement of the 16th series in September quarter 2011. The production of seasonally adjusted individual capital city indexes will be subject to appropriate funding. The unadjusted CPI will continue to be the official headline measure.

6.17 The ABS will publish an All groups excluding food and energy index as part of the analytical series, from the September quarter 2011.


Tradables and non-tradables

Background

6.18 In 1997 the ABS identified a need for some measure of the impact of prices of imported items to determine the extent to which price change is attributable to domestic market factors versus international factors. In September 1999 the Tradables and Non-tradables price indexes were released as part of the CPI analytical series. These indexes decompose the CPI, at the expenditure class level, into items whose prices are largely determined on the world market (tradables), or not (non-tradables).

User views

6.19 The compilation of price indexes that decompose the CPI into tradable and non-tradable components is seen as being of particular use in analysing domestically sourced versus internationally sourced price pressures. The major user of these indexes is the RBA. The RBA is of the view that several of the classifications are out of date and may no longer be valid given changes in the local and world economies.

Evaluation

6.20 The ABS has examined the concept of classifying CPI expenditure classes into tradable and non-tradable components as a suitable consumer price measure, and the approach used to classify expenditure classes as either tradable or non-tradable.

6.21 The tradables and non-tradables price indexes are constructed using a subjective threshold whereby expenditure classes are classified based on the trade exposure of the contributing commodities. The existing methodology will continue to be used in construction of the indexes. All commodities which are classified as importable and/or exportable will form part of the tradable component. The non-tradable component will consist of the remaining commodities.

6.22 The new classifications will be implemented as part of the 16th series CPI in September quarter 2011. The updated classification of CPI expenditure classes will be provided in A Guide to the Consumer Price Index: 16th Series, to be released in October 2011.

6.23 The AG were supportive of the continuation of the tradable and non-tradable indexes and of the ABS decision to update the classifications.

Outcome

6.24 The ABS will update the tradable and non-tradable series classifications, from the September quarter 2011.


Average Retail Prices and the demand for price level data

Background

6.25 The Australian CPI measures price change over time (i.e. a temporal measure) and as such does not provide comparisons between relative price levels at a particular date, either between products or regions (i.e. a spatial measure).

6.26 Average Retail Prices of Selected Items, Eight Capital Cities (cat. no. 6403.0.55.001) provides some price level information. However the Average Retail Prices (ARP) data is limited to providing price levels between the capital cities at a point in time and only for a selected number of items.

User views

6.27 There was limited comment on the usefulness of ARP during the review consultation process. Some suggested that it needed to be expanded to cover more data items.

6.28 A number of users have requested the ABS to provide detailed price level data for analytical purposes, in particular, to compare price levels across localities.

Evaluation

6.29 The ABS is of the view that the ARP data neither performs the function of a temporal measure nor of a detailed price level comparison (spatial measure) in a robust manner. Caveats are placed on the ARP data indicating limitations, but the data continues to be used for purposes beyond its scope. The ABS considers that the ARP data is not fit for the purpose for which it is being used.

6.30 The ABS will assess the demand for price level data more broadly beyond that available in ARP.

6.31 Consideration has been given to the establishment of some form of price level unit record file, i.e. a confidentialised unit record file (CURF). However, further investigation and previous work on this topic has highlighted many practical and legal considerations in releasing price data at the unit record level. The ABS has specific concerns regarding the legal, confidentiality and public interest implications (e.g. possible manipulation of series) of developing a general purpose CPI CURF. This concern, however, must be balanced against the genuine need for price level data for analytical purposes.

Outcome

6.32 The ABS will discontinue the Average Retail Prices (ARP) publication from the September quarter 2011; the June quarter 2011 release will be the final issue.

6.33 The ABS will engage in discussions with potential users of low level CPI data to discuss how their needs might best be met on a case-by-case basis. The ABS will not develop a confidentialised unit record file (CURF) of collected prices.




Previous PageNext Page