RESULTS


INDEX AGGREGATION

The results of analysis to compare the spending patterns of the PBLCI sample population and the CPI sample population based on detailed HES data showed the following. Of the 75 expenditure classes within which the EA weights are not derived from expenditure information specific to the PBLCI households, 52 showed statistically significantly different spending patterns between the two samples. In 24 of these classes, the difference in spending patterns was accompanied by differences in price movements. Further details of this analysis can be found in Appendix A: Expenditure Class Analysis.

The following charts compare the existing all-groups PBLCI index with an experimental version in which the 21 of the 24(footnote 1) expenditure class indexes discussed above are constructed using PBLCI expenditure weights at the EA level. Both indexes have a reference date of March 2006(footnote 2), i.e. the indexes are set to 100 at March 2006.

Figure 2 compares the quarterly percentage changes for the two indexes. The movements show little difference and never differ by more than 0.1 (rounded to one decimal place) percentage points

Figure 3 compares the annual percentage changes for the two indexes. Again, the movements show little difference for the most part with the only perceptible differences at December 2009 (0.1 percentage points, to one decimal place) and March 2010 (0.2 percentage points, to one decimal place)

Figure 4 shows the index levels for the published PBLCI and the experimental series. Both indexes are referenced to March 2006=100. The experimental index tracks slightly above the published throughout and as of March 2011 the levels are 120.5 and 120.3 respectively.

ADDITIONAL ITEMS AND OUTLETS

A further issue the investigation considered is the likelihood of there being items on which the PBLCI households spent a significant amount of their household budget but which were not currently represented in the index. The investigation identified seven such items which could be considered for inclusion in the PBLCI if suitable expenditure information and representative prices can be regularly sourced.

A final consideration was whether the shopping habits of the PBLCI households differed from the main CPI sample population. To enable this analysis, detailed retail receipts were collected from households with the aim of determining when and where they carried out their shopping and whether these habits differed between the PBLCI households and the main CPI households. The information suggested that the shopping habits of the two sets of households did differ for many items and this often led to the PBLCI households paying less on average for those items than the main CPI households.

1 For three of the 24 expenditure classes, it was not possible to complete a price comparison for the full five year period. <back
2 The PBLCI published index is only available from June quarter 2007. Earlier index values have been estimated based on historical 'Age Pensioner' and 'Other Government Transfer Recipient' Indexes for the purposes of this analysis.<back