INTRODUCTION TO A CONSUMER PRICE INDEX
4.1 This Chapter looks more closely at the concept of a consumer price index. (footnote 1) The conceptual basis is important in determining the items that are in scope of the index, their weights and pricing. The previous Chapter provided theoretical background as to how this information is brought together to compile the index numbers.
4.2 As the title suggests, a consumer price index measures the change in prices paid by households for goods and services for consumption purposes. All expenditures by businesses and expenditures by households for investment purposes are out of scope of a consumer price index. In this regard, expenditure on housing presents particular difficulties as it can be considered as part investment and part purchase of shelter-related services.
4.3 While in principle the make-up of a consumer price index seems relatively straightforward, it is less so in practice. A starting point is to address the question of what is the principal purpose of the index. There are three principal purposes identified in this Chapter and the implications of each for index construction are discussed. Other issues are also considered, including the population and the geographic coverage of the index.
DEFINING A CONSUMER PRICE INDEX
4.4 There is currently no single universally accepted definition of a consumer price index. The typically quoted description of such an index is the following statement from the 'Resolution concerning consumer price indices' released in 2003 by the Seventeenth International Conference of Labour Statisticians convened by the Governing Body of the International Labour Organization (ILO):
“The CPI is a current social and economic indicator that is constructed to measure changes over time in the general level of prices of consumer goods and services that households acquire, use or pay for consumption. The index aims to measure the change in consumer prices over time. This may be done by measuring the cost of purchasing a fixed basket of consumer goods and services of constant quality and similar characteristics, with the products in the basket being selected to be representative of households’ expenditure during a year or other specified period.”
4.5 The Resolution effectively suggests three bases on which to compile a consumer price index, namely: the consumer goods and services that were acquired (or taken ownership of); the actual amounts paid (or outlaid) to gain access to consumer goods and services; or the consumer goods and services which were used up (or consumed). All three are measured in the base period. The differences in these conceptual bases will be described later in the Chapter. The approach employed is determined by the principal purpose for which the index will be used.
HOUSEHOLD ACCOUNTS AND A CONSUMER PRICE INDEX
4.6 Consideration of the types of transactions undertaken by households provides a framework for clarifying the domain of a consumer price index.
4.7 Households derive income from various sources including wages and salaries, investment returns and sale of assets. The income received (net of taxes on income) is spent on goods and services, or saved (where saving includes the purchase of assets). Essentially a consumer price index is concerned with the outlays component of the household account. The following table illustrates a household account, showing the types of incomes and outlays essential to the discussion.
4.1 TRANSACTIONS OF THE HOUSEHOLD SECTOR
4.8 The value of the transactions will reflect the actual payments made, or committed to be made, on the purchase of goods and services. Irrespective of whether an item is purchased on credit or paid for in cash, the full value of the item is recorded in the accounts. In an accrual accounting model, the time of recording the transaction is the period in which the change of ownership occurs rather than when the payment is made (in practice they will generally coincide). While most outlays will be valued at market prices, some will be at subsidised prices (e.g. public education, medical services).
4.9 As an index is typically constructed for a group of persons or households, the accounts of those individual persons or households must be added together. There are some results of the consolidation process that should be noted, in particular the treatment of transactions in second-hand goods between households.
4.10 In the accounts of an individual household, the purchase of an existing item (e.g. motor vehicle) from another household would be shown as an outlay. It would also be shown as a negative outlay (i.e. a receipt or ‘income’) in the account of the selling household. (footnote 2) As a result, in the accounts of the household sector as a whole, such a transaction would not be shown - it would ‘net out’ because the expenditure by the purchasing household is negative expenditure by the selling household. However, if such a transaction occurred through a second-hand dealer, then the dealer margin (not the full price of the second-hand good) would be included in the outlays of the household sector since this is expenditure by households that is income in another sector’s accounts.
4.11 The situation becomes more complicated when the transactions are between the household and non-household sectors of the economy. If the second-hand good was previously owned by another sector of the economy, then the full sale price of the good would be shown as an outlay by the household sector. On the other hand, the value of any sale of second-hand goods by the household sector to the non-household sector would be deducted from the outlays of the household sector. Chapter 7 contains a more detailed discussion of the treatment of second-hand goods.
4.12 Another feature of household accounts that generally carries across to consumer price indexes is the treatment of repayments of principal. Regardless of the method of financing any purchase or acquisition, the expenditure that is recorded on an item is its full purchase price. Any repayment of principal associated with borrowed funds to finance the purchase of an item is excluded, because including principal repayments would result in double counting.
PRINCIPAL PURPOSE OF A CONSUMER PRICE INDEX
4.13 A consumer price index may serve various purposes. However, three principal purposes are generally recognised, namely to measure:
|Income and Outlays|
|Wages and salaries|
|Income from assets (interest, dividends, rent)|
|Total available income|
|Private sector goods and services|
|Government goods and services|
|Property purchase (residential, shares, bonds etc.)|
|Interest on debt|
|Savings (balancing item)|
|Property (real and financial)|
Purchasing power of money incomes
4.14 A CPI designed for measuring the purchasing power of money incomes is concerned with answering questions like ‘how much income would be required today to purchase the same basket of goods and services purchased in the base period’. (footnote 3 ) The appropriate domain of the basket in this case is all those outlays on consumer goods and services actually made by households in the base period. In terms of the household account framework, savings (and the use of savings to purchase assets) could be treated as foregone consumption. Thus savings can be excluded from the item coverage.
4.15 In the above context, the correct measure of income on which to base the assessment is net income (after income tax), not gross income. Application of the index to gross income is only valid if income tax is proportional to income and the treatment of property income is identical to that of wage and salary income. A progressive income tax regime such as that applying in Australia emphasises the need to use net income. Further, as the relative significance of different sources of income and expenditure varies considerably between household types, changes in purchasing power are best assessed at a disaggregated level, rather than the ‘all household’ level.
Assessing changes in living standards
4.16 In assessing changes in living standards, the CPI is used in conjunction with data on expenditures by households to measure changes in their volume of consumption of all goods and services.
4.17 An initial issue is to define ‘standard of living’. A narrow definition of standard of living is that it includes the volume of goods and services actually consumed by households in the base period. For many consumer items, the acquisition of, the payment for, and the consumption of the item occur at about the same time. However, for some items the volume of the item consumed in a period may have little relationship to the payments made in the period. Typical examples are durable items (e.g. private motor vehicles, where the vehicle may have been purchased several years earlier), financial charges (e.g. insurance premiums which cover claims payments as well as an insurance service (footnote 4) component; or interest charges, which provide compensation to depositors as well as covering the service costs of the financial intermediary). For other items the price is substantially below the ‘true’ or economic cost of providing the good or service, so that expenditure is not a reflection of the quantity of the item consumed. Typical examples are various services provided by the public sector (e.g. education and medical care). Estimates must be made of the economic value of these items actually consumed in the base period.
4.18 In the case of durables, estimates of the market value of consumption can be made by reference to market prices (such as rents in the case of owner-occupied dwellings and leasing charges for motor vehicles (footnote 5) ).In the case of financial services, estimates of the service component (essentially operating costs plus profit) must be derived from accounts of relevant companies. In the case of public sector goods and services, reference can be made to comparable private sector charges where applicable or estimates can be based on the cost of service provision (e.g. teachers’ salaries plus building and running costs in the case of education).
4.19 A wider definition of living standards could be envisaged. It might include environment related living conditions, such as the quality of air and water, or the area of national parks. While such conditions are important in their own right, the measurement of these factors, the values placed on them by households and the means of including them in an index of living standards are as yet insoluble problems (see Pollak (1989)). For current purposes the narrow definition will be retained.
4.20 Against this background, the domain for an index designed to assess changes in living standards would include:
- changes in the purchasing power of money incomes
- changes in living standards
- price inflation for the household sector.
4.21 This measure accords quite closely to the concept of household final consumption expenditure in the national accounts.
Measuring household inflation
4.22 The objective of the CPI in this case is to obtain a measure of price inflation facing households as consumers. Such a measure is primarily for use in macroeconomic management and also has some possible uses in contracts where an index of prices of household consumption items is appropriate. It should be noted that this is not a measure of economy-wide inflation, as it only relates to household expenditures.
4.23 There is no generally agreed definition of ‘inflation’, which complicates the issue of how it should be measured. Nevertheless, some aspects are suggested. A measure of household inflation should relate to the contemporary rate of change in prices of goods and services. For example, this would exclude interest rates, as interest rates are not prices in the same sense as are prices of particular goods. Rather, interest rates represent the price of money, or more correctly, the relative price of consuming goods and services today rather than in the future.
4.24 An important aspect of an inflation measure is that it should only include ‘actual’, or market-determined prices. Thus an inflation measure would not include imputed rent on owner-occupied dwellings (which however would be included in a cost of use approach, as noted in paragraph 4.36). A measure of house prices would be more appropriate if housing is not considered an investment. Financial assets would not be considered a good or service so that prices of shares and the like would be out of scope. However, such a measure would need to capture changes in the service charges of intermediaries involved in financial and asset transaction services, such as banks, insurance companies and real estate agents.
4.25 It could be argued such an inflation measure should also exclude goods and services provided to households at subsidised prices: the inflation rate has implications for government policy, and as such it should be determined by market forces unhindered by the actions of governments themselves. Subsidies represent distortions of ‘pure’ market forces and subsidised prices do not reflect the true market price (or economic value) of the goods and services. The treatment of taxes and subsidies (which are negative taxes) should be symmetric. However, this would lead to some significant goods and services (for instance, education) being omitted from a CPI that would otherwise be considered essential for complete coverage. Consequently, the most common practice is to include these subsidised goods and services.
4.26 The domain for a measure of household inflation would thus include:
- residential rent payments
- imputed rent on owner-occupied dwellings
- consumer durables
- private sector goods and services (other than financial services and durables)
- the service value of financial services (insurance and banking)
- government-provided goods and services valued at cost or at market prices.
- residential rent payments
- net purchase of owner-occupied dwellings
- net purchase of consumer durables
- goods and services provided at market prices (other than financial services and durables)
COMPARING INDEX OUTCOMES BY PURPOSE OF INDEX
4.27 The intended purpose of a CPI will be the main influence in determining the domain of the index and no single domain can serve all purposes for which a CPI might be used. The following example from Woolford (2001) indicates how the results of various price changes would be reflected in indexes constructed according to the above three purposes:
“Consider the case of a subsidised commodity and a change in government policy that reduces the level of subsidy, all other things being equal. A living standards index would record no change, as there has been no change in the economic price (or value) of the commodity concerned. The money income evaluation measure would record a price increase reflecting the fact that households are now required to pay more out of their own pockets. The ideal inflation CPI would also show no change, as the item would be excluded.
- the value of intermediary services involved in financial and asset transaction services (e.g. insurance, banking and stockbroking services).
Consider the case when the only prices to increase are those for dwelling rents. The living standards index would record a higher rate of increase than either of the other two, reflecting the fact that this index assigns a larger weight to dwelling rents to cover the value of the flow of housing services enjoyed by owner occupiers.
Finally, consider the case of so-called ‘income dependent’ prices, that is prices that depend, at least to some extent, on the income of the purchasing household. Now, assume that the only change in actual transaction prices between two periods can be attributed to changes in household incomes. With no change in the economic value, the living standards index should show no change. The income evaluation index should record the increase. To the extent that this pricing mechanism is generally used to implement a subsidy programme, then an ideal inflation index would exclude the item.”
CONCEPTUAL APPROACHES TO CONSTRUCTING A CPI
4.28 Once the purpose of a CPI has been determined, a method of construction can be identified to satisfy that purpose. Consistent with the three purposes outlined above, there are three alternative conceptual approaches for constructing a CPI. These approaches fall into line with the ILO description of the purposes of a CPI in terms of changes in prices of goods and services whereby "...a reference population acquires, uses or pays for". (footnote 6)
The three methods are:
(i) The Acquisitions method: in the base period, any goods and services acquired (actually received) by the reference population - regardless of the period in which payment or use occurs - are included in the CPI.
(ii) The Cost of Use method: in the base period, any goods and services used (consumed) by the reference population, regardless of when they are paid for or acquired, are included in the CPI. In particular, the cost of using the good or service is measured by its 'true' or 'economic' cost.
(iii) The Outlays method: in the base period, any goods or services for which payments were made to gain access to goods and services (without regard to the source of such funds).
4.29 The acquisitions and outlays approaches are very similar. The acquisitions approach leads to a CPI basket that can be viewed as a subset of the basket resulting from an outlays approach. Both conceptual approaches include goods and services acquired during the base period, while the outlays approach also effectively includes any inescapable costs associated with the acquisition of a good or service, such as interest charges. The cost of use approach can result in a basket that differs from both the acquisitions and outlays approaches.
4.30 The choice of conceptual approach for construction should depend on which approach is most appropriate for the purpose. The approach to construction that is most appropriate for each of the three possible CPI purposes is outlined below:
(i) Measuring household sector inflation. In this case the acquisitions method is the most appropriate. A measure of household sector inflation should relate to the contemporary rate of change in prices and services. The acquisitions approach captures this, by measuring changes in the prices of goods and services actually acquired in the base period.
(ii) Purchasing power of money incomes. In order to determine changes in the purchasing power of money, an outlays approach would be most appropriate. The outlays approach would provide a proxy for household income through measurement of consumer outlays.
(iii) Assessing changes in living standards. The cost of use approach provides the best indication of changes in living standards as it relates to goods and services actually consumed in the base period.
COMPARISON OF THE CONCEPTUAL APPROACHES
4.31 In practice, for most goods and services purchased by the reference population, outlays, acquisition and use all occur within a relatively short space of time and the price paid by the reference population represents a 'true' value, effectively making the distinction unimportant. In some cases however, there can be significant lags between outlays, acquisition, or use or the price paid may differ significantly from what is considered the 'true' value.
4.32 There are three areas of expenditure in which these conceptual approaches provide significantly different results. These are:
(i) purchase of dwellings
(ii) purchase of durable items
(iii) financial services and the use of credit.
4.33 To illustrate the differences, the way in which these three special cases are treated under each approach is outlined below.
Expenditure on housing
4.34 Under the acquisitions approach, the required measure is the change in prices for both the net purchase of housing, and the increase in volume of housing due to renovations and extensions, plus other costs incurred in ensuring the continued supply of services provided by the dwellings (e.g. maintenance costs and council rates) for owner-occupied housing. Changes in rental are measured for that part of the reference population that resides in rented dwellings (costs such as maintenance on these dwellings are paid by investors who are out of scope of a CPI).
4.35 Under the outlays approach, the required measure includes changes in the amount of interest paid on mortgages and the costs incurred in ensuring the continued supply of services provided by the dwellings (e.g. maintenance costs and council rates). Again, also included are changes in the rental, which is measured for that part of the reference population that resides in rented dwellings.
4.36 Under the cost of use approach, the required measure is the change in the economic value of the services provided by dwellings. The price of these services is usually measured as the rental value of the dwellings. In the case of owner-occupied dwellings, the rental values are imputed. Costs such as maintenance costs are not included as they are part of the cost of maintaining an investment and so are outside the scope of a CPI.
4.37 In the case of durable goods, the three approaches result in the following treatments:
(i) Acquisitions - the basket includes those durable goods acquired in the base period and their price measure is the transaction price.
(ii) Outlays - the basket includes those durable goods paid for in the base period and their price measure is the transaction price.
(iii) Cost of use - the basket includes the services of durable goods consumed in the base period, regardless of the period in which they were purchased and the price measure is the market value of the services provided by those goods (measured in business accounts as depreciation plus the return on investment).
Financial services and the use of credit
4.38 Under the acquisitions approach, interest paid is not a charge that is within scope of the CPI basket of goods and services. The service for which prices are required is the service of providing the various banking services (including the provision of loans).
4.39 Under the outlays approach, the product being priced is the cost of servicing all loans taken out to purchase products that are part of the CPI basket. Thus the change in the level of interest paid on a real level of debt is the required price measure.
4.40 The cost of use approach requires that the cost of the financial services used is measured, similar to the acquisitions approach.
4.41 Although these alternative approaches to the construction of a CPI are characterised by conceptual differences, they are more likely to result in short-term rather than long-term differences in outcomes (particularly in the case of the acquisitions and outlays approaches). In the long-term, the outcomes are expected to be similar. In practice, each approach covers a broad range of consumer goods and services, which tend to have similar long-term price behaviour in the absence of external shocks or institutional change. In addition, there is a large set of components common to all three approaches.
CPI ITEM COVERAGE
4.42 The set of goods and services included in a CPI is said to be its item coverage, or more commonly, the CPI 'basket'. In concept, all consumer goods and services are within scope of the index. However, the coverage will vary depending on the principal purpose of the index, as discussed above. Also, if an index is compiled for a subgroup of households, it is possible that their expenditures on some commodities may be nil, but conceptually the item will be in scope.
4.43 There are certain exclusions and inclusions that are relevant to any CPI basket. These are covered below.
Business, savings, and investment related purchases
4.44 As a general principle, a consumer price index only includes consumption goods and services purchased by households. A consumption good or service is basically one from which households directly derive utility or satisfaction. Any business-related purchases are excluded from the basket, as are those items that have a significant savings or investment component, such as land or capital goods. All types of income are also excluded from the basket (except those which directly offset a specific purchase, such as subsidies or trade-ins).
Taxes, levies, concessions & subsidies
4.45 The prices of consumer goods and services and the ability of households to purchase those items are affected by a wide range of government taxes, regulatory processes and assistance measures. The treatment accorded such distortions under the acquisitions and outlays approaches are similar, but there are differences to a cost of use approach.
4.46 As a general principle, the acquisitions and outlays approaches only include government taxes, subsidies etc. whenever such measures are tied to the level of consumption of a specific good or service. Thus any taxes based solely on income will generally be out of scope, whereas the prices of goods and services will be inclusive of indirect taxes and the effects of any commodity specific subsidies. In some cases government taxes and charges may not be directly related to the level of consumption of a good or service but may still be included as they are an inescapable cost of other consumption decisions (e.g. local government rates and charges are an inescapable cost of home ownership).
4.47 A cost of use approach is concerned with the true value of goods and services consumed. For example it will value subsidised items at their comparable market value. It will also exclude income taxes (see Chapter 7 for more information).
4.48 As noted earlier, in concept both the purchases and sales of second-hand goods should be included in a CPI. The purchases of second-hand goods by households are regarded as positive expenditure and sales by households as negative expenditure. The exact treatment of second-hand goods will also depend on the nature and extent of transactions with non-household sectors (see Chapter 7 for more information).
Illegal or ‘undesirable’ goods
4.49 As a general principle, all purchases of goods and services for consumption are in scope of a CPI. They include goods or services that are either illegal or may be considered socially 'undesirable', such as alcohol and tobacco, gambling, prostitution, and illicit drugs. Decisions regarding the composition of the basket are not based on moral grounds.
4.50 All price indexes have a geographic or regional dimension, such as the price index for a specific city or country. A further aspect to the geographic coverage that is important for price collection for a consumer price index is whether the objective is to measure price changes for:
- a set of goods and services obtained in a particular region
4.51 If the index is to measure the prices of items obtained in a region then the ‘basket’ will comprise all consumer goods and services purchased in that region by households for final consumption. These purchases can be made by households that are residents of that region or visitors to the region (including from overseas) and will include purchases by households in other regions made over the Internet or by mail order. If the index is to measure prices of items obtained by household residents of the region, then it will comprise all consumer goods and services purchased by those households regardless of the geographic region from which they were purchased. In addition to purchases made in that region, it will include any purchases these households make whilst visiting other domestic regions or foreign countries as well as items ordered from outside the region over the Internet or by mail order.
4.52 The geographic dimension becomes more important the smaller the region to which the index relates.
CPI reference population
4.53 The expenditures or weights applied to the index basket reflect expenditures of a certain ‘reference population’. Typically the basic unit of reference population is households. (footnote 7) The household is an appropriate unit since all members of the household jointly consume many items, such as food, motor vehicles and housing, and it is not sensible (or possible) to determine an expenditure for each individual.
4.54 A consumer price index can be constructed for all households or a subset of households (e.g. age pensioners, wage and salary earners,
self-funded retirees). Even if the purpose of a CPI requires the broadest possible reference population, certain types of individuals whose consumer expenditures are minimal may be excluded, for example those living in institutions such as hospitals, barracks, prisons, and on board ships.
- a set of goods and services obtained by residents of a particular region. (A region can be any geographic entity, such as rural areas, city, state or country.)
WEIGHTING THE INDEX
4.55 The overall CPI provides a measure of the average rate of price change across all consumer goods and services. In calculating an average of price changes across these items it must be recognised that some are more important than others in terms of their share of household expenditure (or weight). For example, if household expenditure on bread was three times as large as expenditure on cheese, then a 10% increase in the price of bread should have a similar impact on the CPI as a 30% increase in the price of cheese. Thus the aggregate index should be compiled by weighting price movements relative to their importance in the overall expenditure levels of households. The reference period for the expenditure data would depend on the chosen index formula. For instance, a Laspeyres formula would require expenditure data from the initial period while a Paasche formula would require data for the current period.
4.56 Expenditure of the reference population is usually measured over a year to obtain the reference data. Doing this overcomes any problems that might arise from seasonal changes in expenditure (e.g. different types of clothing and foods purchased in summer and winter).
1. Issues are examined from a conceptual perspective. The extent to which concepts can be, or are, implemented in practice is not considered in this Chapter. < Back
2. Households often trade in an existing good to buy the new good. In the individual household account the outlay on the good would be shown as the new good price less the price received for the sale of the old good. < Back
3. This presentation is consistent with an index formulation that uses fixed-base period weights. It is simpler than a cost-of-living index framework. A cost-of-living approach estimates the change in expenditure required to maintain a household at the same level of utility or satisfaction in the reference and comparison periods. For a discussion of the cost-of-living approach see Triplett (1999). < Back
4. See Chapter 7 for an explanation of the insurance service charge. < Back
5. If leasing charges are used to value the volume of service consumed, then certain vehicle charges such as registration and vehicle servicing would not be included separately in the domain as they would be covered by the leasing charges. < Back
6. ILO (1989) page 4. < Back
7. A household is a group of people who usually reside and eat together. This may be a one person or a multi-person household. < Back