6461.0 - Australian Consumer Price Index: Concepts, Sources and Methods, 2005  
ARCHIVED ISSUE Released at 11:30 AM (CANBERRA TIME) 17/08/2005   
   Page tools: Print Print Page Print all pages in this productPrint All  
Contents >> 6. Estimating Weights

INTRODUCTION

6.1 In previous Chapters it has been noted that weighting practices vary at different levels of the CPI. At published levels, weights are reviewed about every five years. Typically these reviews follow the release of data from the Household Expenditure Survey (HES). (footnote 1) At the unpublished levels the weights can be varied at any time. At the elementary aggregate or price sample level there is no explicit use of weights.

6.2 This Chapter describes the procedures that are typically followed in updating the CPI weights.


FIXED LEVEL WEIGHTS

6.3 At the level of the index at which weights are fixed, the ideal is to have a snapshot of all household expenditure. It is important that the data are consistent (expenditures can be added up without worries about coverage, double counting etc.) across the population group and that it relates to the same time period. For example, if expenditures were from different periods for different items, then these could be affected by changes in economic conditions, tax rates and population growth, which could distort the weights. Also, while an information source might provide an estimate of total sales of an item, it would be necessary to know the proportion of those sales to households as consumers. For example, sales of whitegoods and browngoods (footnote 2) will include sales that are not in scope of the CPI, i.e. to the business sector and households as owners of rental properties.

6.4 The HES is the only source of data that comes close to the ideal. However, while it provides a comprehensive coverage of household expenditures, various adjustments need to be made to the HES data for use in the CPI, and for some items it is supplemented or replaced by other data sources.

6.5 The weaknesses in the HES data for CPI purposes include:

  • The HES is a survey and is thus subject to sampling error. It is possible for the selection of one or several households with exceptionally high expenditure on a certain item/s to significantly affect the expenditure estimate for that item at a regional level, especially in the smaller capital cities.
  • Some expenditures recorded in the HES are not fully contemporaneous. Estimates for some items which are more expensive or purchased infrequently are obtained or supplemented by recall, rather than relying on expenditures actually recorded in the diary during the two week interval the household was included in the survey. Periods over which households are asked to recall expenditure varies by item - it can be the period of the last payment or up to two years in the case of house purchase and sale. No adjustments are made to HES data for any change in prices between the period of recall and that of recording in the HES.
  • The HES records all data as reported by households, with no adjustment for known cases of underreporting, especially in expenditure on alcohol and tobacco.
  • Some expenditures captured in the HES are not fully within the period of recall and that of recording in the HES.
  • There are some expenditures required by the CPI which cannot be obtained from households, such as insurance services.

Determining the CPI structure

6.6 An initial starting point is to determine a CPI structure and item coverage that are consistent with the principal purpose and conceptual base and which take account of recent and expected market trends. For example, changes were made to the structure of the CPI for the 13th series to reflect the change in principal purpose to a measure of inflation. Further changes were made in the 14th series in order to better accommodate possible changes in expenditure patterns as a consequence of The New Tax System. For example, in the 13th series CPI, photographic equipment was included with Audio, visual and computing equipment in order to better allow for the growing use of digital photography (it is both complementary and competitive with other visual equipment). In the 14th series, the repair of appliances was included with purchase to reflect the substitution being made between repairing and purchasing new appliances.

6.7 As previously noted, a prime consideration in deriving the CPI structure is assessing the utility that households derive from commodities and the extent of substitution between commodities. Another consideration is the need to balance the relative importance of those commodities in household expenditure with the reliability of estimates of those expenditures. For example it would make little sense to have an expenditure class for large durable recreation vehicles (say speed boats, dune buggies) as such purchases identified in the HES will be small in number (less so in value) and volatile over time.

6.8 The commodity classification used in the CPI is based on the international standard classification of Classification of Individual Consumption by Purpose (COICOP). This classification is based on the concept of household utility. A significant advantage of adopting a COICOP-based classification is to align the CPI more closely with the dissection of household final consumption expenditure in the Australian national accounts (also based on COICOP).

HEC/CPI concordance

6.9 Expenditures recorded in the HES were coded according to the Household Expenditure Classification (HEC) (footnote 3) . In order to derive expenditures for the CPI expenditure classes a concordance was established with the HEC codes at their most detailed level (10 digit code). Establishing the concordance involved examining detailed listings of commodities coded to each HEC code. The concordance is available as an Excel spreadsheet in Consumer Price Index: Concordance with Household Expenditure Classification, Australia (cat. no. 6446.0.55.001 available on the ABS website).

6.10 A majority of HEC codes could be exclusively allotted to a CPI expenditure class. For example all of HEC 0302019902, ‘Smallgoods’ was allotted to the CPI expenditure class Other fresh and processed meat. However there are numerous HEC codes where a one-to-one correspondence could not be established (there are just over 600 HECs at the 10 digit level and only 89 CPI expenditure classes). The reasons why unique concordances could not be established include:
  • The HEC code may not be sufficiently detailed. For example HEC 0302019901, ‘Mince’ could be prepared from a variety of meats. For CPI purposes it was decided that mince is not processed meat and hence expenditure on mince was spread across appropriate CPI meat expenditure classes.
  • Information provided by households does not allow a commodity to be clearly identified. These expenditures are reported in codes such as 0302000000, Meat (excluding fish and seafood) nfd’ (where n.f.d. is an abbreviation for ‘not further defined’). Again these expenditures were spread across appropriate CPI expenditure classes.
  • Households cannot or do not separately identify some expenditures. For example, in some states, state governments operate compulsory third party vehicle insurance schemes, and the amount of insurance paid is included with the vehicle registration charge, often resulting in households reporting the combined amount only. In this case a split was derived from average registration and insurance charges collected for the CPI.

6.11 Where HEC codes were split across expenditure classes, the splits were determined pro rata, using any industry or other data sources available or subjectively. In most cases the expenditures involved were relatively small.

Adjustments for under reporting

6.12 There are non-HES data sources that provide accurate estimates of household expenditure on some commodities (often only at the national level). These alternative estimates are used for validating the HES data and, for most products included in the CPI, were in accordance with HES estimates. However, there were significant differences between HES and other data sources in the estimates for alcohol and tobacco expenditure. For both these product groups the major consumer is households and the imposition of excise and other taxes on these products means that non-HES estimates are more accurate than the understated estimates available from the HES due, mainly, to regulatory reporting requirements.

6.13 For alcohol and tobacco, under-reporting factors were derived at the national level using household final consumption expenditure estimates from the national accounts. These factors were then applied to the HES expenditure estimates for each capital city. The 1998-99 HES has tobacco expenditure at around two-thirds and alcohol expenditure at a little over half of the respective national accounts estimates. The same adjustment factor was applied across all alcohol expenditure classes.

6.14 The intention of the CPI classification is that all alcohol expenditure be in the alcohol subgroup. As noted earlier, households do not always separately identify expenditures as fully as ideally required. In the case of restaurant meals, expenditure on alcohol is not always reported separately to the meal component and is sometimes reported simply as ‘drinks’. A sample of household records was examined and, by applying various assumptions, an estimate was made of the proportion of reported restaurant meal expenditure that was alcohol and this factor was applied across each capital city.

Recall adjustment

6.15 As noted above, some expenditures were collected for HES on the basis of recall rather than as entries in the diary over the two weeks for which the household was in the survey. The extent of the recall period varies: for some items it is purchases in the last three months (most white goods and furniture, house repairs); last payment (general rates, electricity, health services); last 12 months (most vehicle related expenditures, education, overseas travel, house alterations and additions), and the last two years (house purchase). To the extent that prices for these items change between the time the household last purchased or paid for the item and their inclusion in the HES, expenditures will not reflect completely accurately the underlying quantities acquired during 1998-99. Adjustments are made to some items to account for the price changes.

6.16 No adjustments are made for any items where the recall period is the last three months since any adjustments to be made would be relatively minor. Most items where the recall period is the last payment are typically items where the billing cycle is quarterly or less or where there are options to pay periodically (e.g. local government rates). This leaves only a small number of items such as vehicle purchase, general insurances, overseas holiday travel and education where expenditures could have been incurred up to 12 months prior to inclusion in the HES. Adjustments are made for these items.

6.17 The adjustment procedure for a recall period of 12 months is as follows. In the first quarter (Q1) of the HES reference year the period potentially covered will be from the start of the corresponding quarter of the previous year (for those households selected in the first week of the HES reference year) through to the end of the first quarter of the reference year. Similarly, the pricing period for the households first included in the last week of Q1 in the reference year will commence in the last week of Q1 of the previous year. If expenditures are distributed uniformly over the period, then effectively expenditures in the reference quarter and the corresponding quarter of the previous year will only be half that for the intervening quarters. Thus assigning weights of say one for the intervening quarters and 0.5 for the start and end quarters, we can derive a weighting pattern as shown below.


Graphic representation of weighting pattern derivation



6.18 Thus the adjustment factor for items with a 12 month recall for the 1998-99 HES is:

Formula for an recall adjustment factor calculation


where IS98 is the CPI index number for the expenditure class for the September quarter 1998 etc.

Using the following hypothetical index numbers:

IS97=99; ID97=100; IM98=101; IJ98=102; IS98=103; ID98=104;

IM99=105; IJ99=106


The result of the above formula is:

      (103+104+105+106) x 0.25
_____________________________________________________________
(.03x99+.09x100+.16x101+.22x102+.22x103+.16x104+.09x105+.03x106)


= 1.02 (which is the adjustment factor to be applied to the recalled price).


6.19 With a generally low rate of price change over 1997-98 and 1998-99, the adjustments made for recall were small, although quite significant adjustments have been made in the past (e.g. for the 1984-85 HES).

Aberrant expenditures

6.20 Expenditure estimates for the CPI weights can be validated in time and spatial dimensions. Validation over time requires expenditure to be on a common pricing base (e.g. revaluing 1993-94 HES expenditures to 1998-99 prices and comparing with the 1998-99 HES results will show changes in volume terms). Any large differences can thus be investigated to see if they are valid. For example there was a large rise in expenditures on motor vehicles (both in nominal and constant prices) between the 1993-94 and 1998-99 HESs but this is broadly consistent with new vehicle sales.

6.21 The spatial dimension involves comparing expenditures across the capital cities. Estimates for the smaller capitals are obviously subject to larger sampling errors and thus expenditure estimates for some items of expenditure can be affected by what might be considered atypical expenditure patterns of a small number of households. Large differences were identified between cities in expenditures in terms of average weekly expenditure per household or proportions of total expenditure. They were subject to further investigation and adjustments were made if the differences could not be attributed to factors such as climate or location. A small number of such adjustments were made in the smaller capital cities.

Expenditures not sourced from HES

6.22 For the 14th series CPI it was not possible to obtain suitable data from the HES for three items, namely new house purchase, tertiary education and general insurance services.

6.23 The taget domain for the CPI is all consumer goods and services acquired by households. With respect to owner-occupied housing (footnote 4), the weight accorded to it in the index should only reflect net acquisitions by the household sector and expansion of the housing volume through alterations and additions (since the CPI measures acquisitions by households in aggregate). Sales of houses that take place between households should be excluded, so that the weights relate only to net additions to the housing stock arising from purchases from outside the household sector (e.g. from businesses such as builders and developers).

6.24 An additional factor that arises with housing is that houses, as well as providing shelter services to the owners, also include an investment element as houses typically appreciate over time. For the purposes of the CPI, the costs of housing should reflect only the service (or consumption) component. A common approach is to regard the cost of the land as representing the investment component and the cost of the structure as representing the consumption component. For the purposes of the CPI, the land component needs to be excluded from expenditure on housing.

6.25 The HES data on alterations and additions was used for that component, but the HES data were considered inadequate for net purchases of dwellings. Even with a two-year recall period the number of house transactions recorded was too small to produce reliable results for each individual capital city. In addition, the values reported in the HES included the land component. An alternative approach of separately estimating the number of house purchases and applying an average house value excluding land had to be adopted.

6.26 Estimates of the net change in the number of owner-occupier households can be derived from the Census of Population and Housing, the HES and the Survey of Housing Occupancy and Costs. After some investigation, the preferred estimates for each capital city were obtained by taking the average of the increase in the number of households between 1996 and 1999 calculated by applying the projected growth rate of households from Household and Family Projections Australia (cat. no. 3236.0) to the 1996 census estimates of the number of owner-occupier households. Estimates of the average value of private dwelling completions by capital city for 1998-99 were obtained from Building Activity, Australia (cat. no. 8752.0). The average value estimates for completions relate only to the cost of construction and therefore enable the land component in house prices to be excluded from the weighting base.

6.27 The CPI is required to measure the change over time in existing dwelling structures, newly constructed dwellings and alterations and additions. The appropriate price for existing dwellings is current replacement cost. Therefore, the prices of newly constructed houses can be used to measure changes in the prices of existing houses.

6.28 In Chapter 7 the measurement of general insurance for the CPI is presented in detail. Briefly, the measurement and weight for general insurance reflect the service provided by insurers rather than the gross premiums paid by policyholders. The expenditures on products funded by insurance claims are included within the expenditures for those products in the relevant area of the CPI. The value of the service of general insurance cannot be collected in the HES. Rather, estimates of the proportion of gross premiums that represents the insurance service are derived from information collected by the Australian Prudential Regulation Authority (APRA) and insurance companies. These are applied to the HES estimates of expenditure on gross premiums for general insurance. In the case of motor vehicle write-offs and vehicle smash repairs the HES estimates exclude expenditure funded from insurance claims. Again the data from APRA and insurance companies provide an estimate of claims payments relative to premiums that is added to those items.

6.29 The weight for medical insurance, on the other hand, does involve the gross premiums. For general insurance, expenditure on specific products funded by claims is included in the CPI separately from the insurance service component of premiums. In the case of medical insurance, expenditures on individual types of medical services are not priced; rather the gross premiums paid are considered are a suitable measurement of payments for all medical services covered (including the insurance service component) and no redistribution of expenditure to individual medical services is carried out.

6.30 Another area where the HES expenditures are not consistent with the conceptual basis for the CPI is in regard to expenditure on university education. Under an acquisitions approach such expenditures should reflect paying the full fee upfront (the full amount if expenditure is deferred under the Higher Education Loan Programme (HELP) or the discounted fee if payment is made immediately). The HES records payments made in full plus any HELP repayments made through the tax system. HELP repayments expenditure is replaced by an estimate of the full fee for deferred payments based on various data sources including the Annual Report of the Department of Education, Training and Youth Affairs.

Revaluing expenditures to the link period

6.31 The expenditure weights derived from HES are based on expenditures in 1998-99. This new expenditure pattern was not introduced into the CPI until the June quarter 2000, the link period. As the quantities underlying these expenditures had to be preserved, the expenditures were revalued to June quarter 2000 prices. This step simply involves multiplying the 1998-99 expenditures at each expenditure class by the ratio of its price index for the June quarter 2000 to the average of its quarterly indexes for 1998-99.

Adjustments for quantity shifts

6.32 Ideally the CPI weights should be as up-to-date as possible and be broadly representative of the expenditure pattern that might be expected over the life of the index series. Thus, when the June 2000 quarter link was being introduced, it was necessary to consider whether any developments and policy changes since 1998-99 might significantly affect the expenditure pattern and whether any revalued expenditures needed to be adjusted.

6.33 A major policy change over this period was the introduction of Lifetime Health Cover. This policy contributed to a substantial increase in the proportion of the population covered by private health insurance. An adjustment was made to allow for the rise in private health insurance based on the information available at the time. (footnote 5)

6.34 Other items where expenditures were likely to have changed between 1998-99 and June 2000 were also investigated. These included the purchase of computers, Internet usage and road tolls. Some adjustments were made, although the effect on the weights at the expenditure class level was minimal.

6.35 In principle, adjustments for any significant change in expenditures between the HES collection time and inclusion of the weights into the CPI is appropriate. However, this raises a methodological issue. The adjustments to expenditures are generally made without compensating adjustments to other expenditures in the CPI basket. In other words it is implicitly assumed that increased expenditure on health insurance, for example, comes from savings and not from reductions in expenditure on other items.


LOWER LEVEL WEIGHTS

6.36 The weights and structure of the CPI below the expenditure class level are continuously subject to review and may be varied at any time. These changes are made through a process termed ‘sample review’. Essentially a sample review involves selecting a component of the CPI (it could be one or more expenditure classes, or part of an expenditure class) and subjecting it to detailed examination. The review determines what changes should be made to the items priced, the outlets they are sourced from and the weights to be applied to the commodities and outlets. For example, it might be determined from recent data that it is appropriate to introduce a price sample under the Fats and oils expenditure class for products composed of both butter and margarine or to change the relative importance of large retail outlets and convenience stores in the price samples for soft drinks.

6.37 Sample reviews are undertaken on a scheduled basis or in response to market changes identified. In general, the aim is to review all components of the CPI over the life of a CPI series (i.e. roughly once every five years).

6.38 Information from any reliable source is used to assess the relative importance of commodities. Sources include data collections of the ABS and other organisations, and publications by industry organisations. Information from the HES is also considered but, for the main part, is not sufficiently detailed or reliable at the lower levels of the CPI structure. For example, the HES data for types of appliance purchased would not be as reliable as industry sales data due to the relatively small samples in the HES.

6.39 At the price sample or elementary aggregate level there are no explicit weights. Rather the price samples are constructed so as to be self-weighting. For example if there were a price sample for medium (100 g) chocolate bars and the major grocery outlets had 80 per cent of such sales and vending machines 20 per cent, then the price sample would be selected so that for every price from a vending machine there were 4 prices from the major grocery outlets.

Footnotes

1. For information on the conduct of the HES, refer to Household Expenditure Survey Australia User Guide 1998–99, cat. no. 6527.0. < Back

2. Browngoods include household electrical items such as TVs, DVD players, sound systems, kettles and toasters. < Back

3. Available in Household Expenditure Survey, Australia: User Guide, 1998-99 (cat. no. 6527.0). < Back

4. The restriction to owner-occupied dwellings reflects the CPI being confined to expenditure on consumer goods and services (renting households have their expenditures on rent included in the CPI while owner-occupiers do not have any rents imputed in the CPI). < Back

5. An adjustment was made for the quantity change only as the price index for private health insurance already captured the impact on premiums of the Federal Government rebate. < Back



Previous PageNext Page