1350.0 - Australian Economic Indicators, Jan 2000  
ARCHIVED ISSUE Released at 11:30 AM (CANBERRA TIME) 10/01/2000   
   Page tools: Print Print Page
  • About this Release
  • Experimental Price Indexes for Age Pension Households: an Update (Feature Article)

Feature Article - Experimental Price Indexes for Age Pension Households: an Update

INTRODUCTION

With the introduction of the 13th Series in respect of September quarter 1998, the Australian Consumer Price Index (CPI) has been designed specifically to provide a general measure of price inflation for all private households in the eight capital cities. Prior to this, the CPI was designed to measure changes in the living costs of capital city wage and salary earner households only.

The historical limitation of the population coverage of the CPI often led to questioning of how well the CPI captured changes in the living costs of other household types.

This article presents an experimental price index constructed to measure changes in the living costs of age pensioner households and compares this with the official CPI. The comparison spans a period of almost eighteen years - from September quarter 1980 to June quarter 1998. It updates and improves upon previous studies released in 1992 (Information Paper The Australian Consumer Price Index: Feasibility of Constructing Price Indexes for Special Population Groups, ABS Cat. No. 6445.0) and in the July 1993 and September 1995 editions of Australian Economic Indicators (ABS Cat. No. 1350.0).

As announced at the time of the introduction of the 13th Series CPI, the ABS plans to develop analytical indexes specifically designed to measure changes in the living costs for a number of subgroups in the population. These new indexes are expected to be published for the first time during 2000.


ASSUMPTIONS AND METHODOLOGY

Construction of a price index of this type requires two inputs: item prices (or measures of price change) and a set of item weights (or measures of relative importance of items).

In constructing the experimental index it has been assumed that the CPI price samples provide measures of price change at the CPI expenditure class level for each capital city relevant to age pensioner households. Although age pensioners can face different prices than the population at large - largely due to age pensioner discounts for government provided services such as public transport fares - the movement in these prices is still likely to be similar to the movement in the non-discounted prices.

Weights for the age pensioner index were calculated as follows: average weekly expenditure data for age pensioner households in each capital city were obtained from the 1988-89 Household Expenditure Survey (HES). These expenditures were aggregated to the expenditure classes used in the 12th series CPI (from June quarter 1992). Adjustments were made for under reporting of spending on tobacco and alcohol, and expenditure aggregates were adjusted to June quarter 1992 prices using movements in the relevant CPI capital city expenditure class indexes. The resulting average weekly household expenditures were converted to weight form by expressing expenditure on each expenditure class as a percentage of total expenditure.

This update of the age pensioner index differs slightly from the previous update in that, from December quarter 1986, indexes have been constructed at the individual capital city level and then weighted together to produce the index for the eight capital cities. This index has been linked to the previously compiled age pensioner index at December quarter 1986 to produce a continuous series from September quarter 1980 to June quarter 1998.

For the purpose of deriving expenditure data from the HES, age pensioner households were defined as capital city households with the age pension as the largest source of income and weekly household income in 1988-89 of less than $175 in the case of one person households or $290 in the case of two person households.

To facilitate longer-term comparison, the experimental age pensioner index and the CPI were re-referenced to a common reference base of September quarter 1980=100.0. These indexes are presented at Appendix 1.


WEIGHTING PATTERNS

Table 1 shows average weekly expenditures at the major commodity group level, for age pensioner households and wage and salary earner households at June quarter 1992 prices. (The selection of a June quarter 1992 pricing basis facilitates comparison with published information on the official 12th series CPI which was linked at June quarter 1992.)

TABLE 1: ESTIMATED AVERAGE WEEKLY EXPENDITURE BY MAJOR COMMODITY GROUP AT JUNE QUARTER 1992 PRICES


Expenditure category
Experimental age pensioner index
CPI

$
$
Food
53.17
125.02
Clothing
14.51
42.74
Housing
30.93
108.48
Household equipment and operation
44.20
125.34
Transportation
24.65
108.94
Tobacco and alcohol
12.78
51.00
Health and personal care
20.48
46.74
Recreation and Education
15.85
74.03
Total
216.57
682.29

(a) As defined for hte purpose of constructing the 12th series CPI, ie, those capital city households that obtain at least three-quarters of their total income from wages and salaries excluding the top 10 per cent (in terms of income) of such households.


Table 1 shows that there are significant differences in absolute expenditures between the two population groups, with wage and salary earner households spending significantly more than age pensioner households on all major commodity groups. However, in studying the behaviour of alternative price indexes, it is not the differences in absolute expenditures that are important but the differences in the proportion of total expenditure accounted for by each category; this is reflected in the weights. Table 2 presents the expenditure weights used to construct the age pensioner index and the equivalent weights used to construct the 12th series CPI.
TABLE 2: EXPENDITURE WEIGHTS BY MAJOR COMMODITY GROUP AT JUNE QUARTER 1992 PRICES


Expenditure category
Experimental age pensioner index
CPI

%
%
Food
24.55
18.32
Clothing
6.70
6.26
Housing
14.28
15.90
Household equipment and operation
20.41
18.37
Transportation
11.38
15.97
Tobacco and alcohol
5.90
7.48
Health and personal care
9.46
6.85
Recreation and Education
7.32
10.85
Total
100.0
100.0


Table 2 shows notable differences in weighting patterns. It illustrates that age pensioner households allocated a higher proportion of their expenditures to food, clothing, household equipment and operation and health and personal care commodities than did the reference CPI population. A higher share of wage and salary earner households’ expenditures went to housing, transportation, tobacco and alcohol and recreation and education. Variations in the expenditure patterns reflect income and demographic differences between the two population groups. For example wage and salary earner households on average have a higher disposable income than age pensioner households, are more likely to have dependent children and have a higher incidence of motor vehicle ownership. On the other hand, age pensioner households have a higher incidence of outright home ownership (ie are less likely to have a mortgage).

TABLE 3: EXPENDITURE WEIGHTS FOR SELECTED EXPENDITURE CLASSES AT JUNE QUARTER 1992 PRICES


Expenditure category
Experimental age pensioner index
CPI

%
%
Take away foods
1.32
3.14
Mortgage interest charges
0.10
6.61
Consumer credit charges
0.57
2.50
Local government rates and charges
4.23
2.19
House repairs and maintenance
4.60
1.83
Electricity
3.10
1.75
Motor vehicles
1.88
3.99
Pharmaceuticals
2.15
0.82
Education fees
0.01
1.56



Differences in expenditure patterns are also noticeable at the expenditure class level. Weights for selected expenditure classes where there are substantial differences between the two population groups are shown in Table 3. These tend to further highlight differences in the income and demographics of the two populations.


OUTCOMES AND DISCUSSION

Chart 1 shows index levels for both the experimental index for age pensioner households and the CPI from September quarter 1980 to June quarter 1998. Despite quite marked differences in weighting patterns, the chart illustrates that the differences in index outcomes were almost imperceptible. Over the whole period, the age pension index rose by 155.5% while the CPI increased by 153.1% - a relatively small difference over an
eighteen year period.
CHART 1: INDEX NUMBERS
Chart 1: Index Numbers


While the change in each index over the entire period of the study is similar, differences in behaviour do emerge over shorter comparison periods. These differences are most readily identified by comparing annual percentage rates of change in the two series (constructed as the percentage change from the corresponding quarter of the previous year) as shown in Chart 2.
CHART 2: PERCENTAGE CHANGE
(from corresponding quarter of previous year)
Chart 2: Percentage change from corresponding quarter of previous year


There are two particularly distinct periods identifiable from the above chart. The first, running up to and including December quarter 1986, is characterised by the behaviour of the two series being almost indistinguishable; while the period from March quarter 1987 is characterised by discernible but not consistent
differences. It is no coincidence that March quarter 1987 marks a change in the relative behaviour of the two series - this represents the quarter that mortgage interest and consumer credit charges were first introduced to the CPI.

This illustrates the point that the extent to which differences in expenditure patterns influence the behaviour of price indexes depends on the variation in individual price movements. If prices of all items move at exactly the same rate then differences in expenditure patterns are irrelevant. The more divergent price movements become, the greater the role of expenditure patterns in measuring the aggregate impact of price change.

Movements in interest rates are often large and against the trend of current price change. Therefore, differences in the relative importance of interest charges across population groups are likely to result in differences in measures of the change in living costs. The observable differences between the two index series post December quarter 1986 are largely attributable to mortgage interest and consumer credit charges which have a combined weight in the age pensioner index of only 0.67% compared with 9.11% in the CPI (see Table 3 above).


CONCLUSION

This study confirms that price indexes constructed to measure changes in living costs for different household types are unlikely to differ significantly over periods of time spanning and extending beyond the business cycle. Over shorter time periods, however, significant differences in index behaviour may emerge.

The inclusion of mortgage interest and consumer credit charges in the CPI from March quarter 1987 resulted in a CPI better suited to its then principal purpose of measuring changes in the living costs of employee households. A consequence was that the CPI became a less suitable measure of shorter-term changes in the living costs of other household types.

As a result of a review of the CPI undertaken during 1997, the ABS decided to change the principal objective of the CPI to one of measuring price inflation for the household sector as a whole. This resulted in some significant design changes being introduced from September quarter 1998. The most noticeable of these changes was the exclusion of mortgage interest and consumer credit charges from the index and the inclusion of net expenditure on the acquisition of dwellings (excluding land). A consequence of this most recent change is that, despite not being specifically designed for this purpose, the CPI is now assessed as providing a better measure of changes in the living costs of age pensioner households.


APPENDIX
APPENDIX 1: EXPERIMENTAL PRICE INDEX FOR AGE PENSIONER HOUSEHOLDS AND THE CPI (a)

Quarter
Expenrimental age pensioner index
CPI
Quarter
Experimental age pensioner index
CPI

1980
Septermber
100.0
100.0
1990
March
202.8
211.1
December
102.3
102.1
June
206.4
214.4
1981
March
104.7
104.6
September
207.6
216.1
June
106.7
106.9
December
212.7
221.8
September
109.3
109.0
1991
March
214.3
221.3
December
114.1
113.6
June
214.5
221.8
1982
March
116.1
115.7
September
216.7
223.0
June
118.8
118.4
December
218.7
225.1
September
122.8
122.6
1992
March
219.9
225.1
December
126.4
126.2
June
219.7
224.5
1983
March
129.1
128.9
September
220.3
224.7
June
132.2
131.6
December
222.1
225.7
September
134.1
133.9
1993
March
225.2
227.8
December
136.9
137.0
June
225.8
228.7
1984
March
136.2
136.4
September
227.2
229.7
June
136.1
136.8
December
228.6
230.1
September
138.1
138.5
1994
March
230.2
231.0
December
139.8
140.6
June
231.2
232.6
1985
March
141.9
142.5
September
233.3
234.1
June
145.0
145.8
December
234.7
236.0
September
148.1
149.2
1995
March
237.5
240.0
December
151.3
152.1
June
239.3
243.1
1986
March
154.6
155.6
September
242.3
246.0
June
157.5
158.2
December
244.0
247.9
September
161.2
162.3
1996
March
244.4
249.0
December
165.4
166.9
June
245.6
250.6
1987
March
168.0
170.3
September
247.2
251.3
June
169.6
172.8
December
249.2
251.7
September
172.9
175.7
1997
March
251.0
252.1
December
175.5
178.9
June
251.8
251.5
1988
March
178.7
182.0
September
251.4
250.4
June
182.2
185.1
December
253.0
251.0
September
185.8
188.7
1998
March
253.7
251.7
December
188.4
192.5
June
255.5
253.1
1989
March
191.0
194.4
June
194.7
199.2
September
197.9
203.8
December
200.5

(a) Base: September quarter 1980 = 100.00.