The 2020 annual re-weight of the Australian Consumer Price Index

This information paper presents the 2020 annual re-weight of the Consumer Price Index (CPI) and Selected Living Cost Indexes (SLCIs). The 2020 re-weight will apply from the December 2020 quarter for the CPI, which will be released on 27 January 2021. The SLCIs will be released on 3 February 2021.

The ABS has annually re-weighted the CPI since 2018, predominantly using Household Final Consumption Expenditure (HFCE) data, which is lagged by 18 months. Annually re-weighting the CPI ensures that the CPI basket continues to be representative of spending by Australian households. Further details on annually re-weighting the CPI can be found in the information paper An Implementation Plan to Annually Re-weight the Australian CPI, 2017.

Normally household spending patterns change slowly. The COVID-19 pandemic has caused a sudden, and in some cases, sustained change in household spending patterns. To reflect this, the 2020 annual re-weight uses a range of more timely data sources (e.g. Retail Trade data and scanner data) for approximately 20 per cent of the weight of the CPI.  This ensures that the underlying weighting patterns for the CPI continue to be reflective of spending by Australian households. To guide the re-weight decision-making process, the ABS developed a framework, which has been used to guide the use of more timely data sources.

For 80 per cent of the CPI, the updated weights were derived in the usual way outlined by the 2017 information paper. For the 20 per cent of the CPI where more timely data sources were used, it was for CPI components such as food, alcohol, electricity and domestic and international travel – for more details see appendix 1.

The ABS will continue to monitor expenditure throughout 2021 to ensure that the CPI weights are reflective of spending patterns by Australian households.

This information paper provides an overview of the changes to the CPI and SLCIs that will be introduced with the 2020 re-weight and presents the updated weighting patterns.

For further information contact prices.statistics@abs.gov.au

Since the introduction of the CPI in 1960, the ABS has periodically reviewed the CPI to ensure that it continues to meet community needs. Traditionally, these reviews have coincided with the release of the ABS’s Household Expenditure Survey (HES), which is conducted every six years. In 2018, the ABS moved to annually re-weighting the CPI using Household Final Consumption Expenditure (HFCE) data to update the CPI expenditure weights. Further details are available in the information paper Increasing the Frequency of CPI Expenditure Class Weight Updates.

HFCE data is a high-quality source to update the CPI weights. It is simple to align to both the CPI concepts and the components measured in the CPI. HFCE data is available for use with a lag of 18 months after the period the data represents. In the case of the 2020 re-weight, HFCE data is available for the 2018-19 financial year.

Typically, changes in consumer spending patterns occurs gradually (see figure 1). Therefore, the 18-month lag is a reasonable amount of time to accurately reflect contemporary spending by Australian households.

However, in 2020 Australian households responded to the COVID-19 lockdowns, business restrictions and social distancing, with significant changes in their spending across a range of goods and services. Figure 2 shows the sudden and significant change in retail trade activity following the outbreak of the pandemic.

To ensure an accurate measure of inflation, the CPI expenditure weights need to capture recent changes in spending by Australian households. In response, the ABS has made use of more timely data to supplement the 2018-19 HFCE data.

Figure 2 shows how consumer spending patterns for a range of goods and services changed significantly in response to the COVID-19 pandemic. In some cases, expenditure returned to pre-COVID levels. For example, clothing expenditure was considerably lower in mid-2020, but has since returned to pre-COVID levels. In other cases, the change in expenditure has been sustained and spending remains significantly above or below 2019 levels, such as spending at cafes and restaurants.

The sustained change in spending is underpinned by a number of real-world changes brought about by COVID-19, for example:

• Increased working from home;
• Domestic and international travel restrictions;
• Social distancing requirements; and
• Changed economic landscape e.g. higher unemployment

The ABS has developed a framework to guide decision making of when it’s appropriate to use more timely data to update the weight.

The ABS adopted this framework to guide decisions on the most appropriate data source to re-weight each component of the CPI. The framework was used for the following scenarios:

1. If spending returned close to pre-COVID levels, 2018-19 HFCE data would continue to be used.
2. If spending hadn’t returned close to pre-COVID levels, a determination was made using more timely data as to whether the change in spending was temporary or permanent.

Where the change in spending was temporary, 2018-19 HFCE data has been used. Where the change in spending was more permanent, data available from the September 2020 quarter has been used.

The outcome from the use of this framework are weights which reflect contemporary spending patterns and ensures that the CPI remains an accurate measure of inflation.

The weights used in the CPI represent annual consumer spending. In 2020, the largest change in weight was for international holiday travel, following the closure of Australia’s borders. For the 2019 re-weight, international holiday travel had a weight of 3.38%, which represented about $18.7 billion in spending by Australian households. Following the 2020 re-weight, the updated weight is 0.08%, representing about $0.4 billion of spending.

If 2018-19 HFCE data was used to update the weight for international travel, its weight in the CPI would have been 3.43%, representing $19.8 billion of spending in 2021. With Australia’s borders remaining closed, this would have led to international travel having too high a weight in the CPI and the CPI basket being unrepresentative of spending by Australian households.

With the uncertainty around the future impacts of COVID-19, the ABS will continue to monitor the spending patterns of Australian households throughout 2021. As part of the 2020 re-weight, the ABS investigated a range of possible scenarios for spending patterns and developed a set of CPI weights for each scenario. Appendix 2 lists the re-weight approach used for each scenario.

Should spending patterns unexpectedly and significantly change, the ABS is well positioned to respond. Where possible the ABS will consult extensively with users and announce any response prior to their implementation.

Every six years, the Household Expenditure Survey (HES) is used to provide information on the expenditure weights of Australian households.  In 2018, the ABS moved to annually re-weighting the CPI using National Accounts Household Final Consumption Expenditure (HFCE) data in the intervening HES years. HES expenditure data is next available in 2023.

The main data source for the 2020 CPI and SLCI weight update is 2018-19 HFCE data. Conceptually, the measurement of HFCE closely aligns with the HES. HFCE data captures household expenditure by Australian households only and excludes expenditure by non-residents in Australia.

While HFCE is the main data source for updating the CPI weights for the inter-HES years, weights for some of the components of the CPI and SLCIs are derived from alternative data sources. These components are:

• New dwelling purchase by owner-occupiers;

• Other financial services; and
• Mortgage interest charges (only included in the SLCIs).
   

For more information on the data sources used for these components see the information papers: Introduction of the 17th Series Australian Consumer Price Index and An Update on the Annual Re-weighting of the Australian CPI and Living Cost Indexes.

For CPI components where expenditure has returned to within five per cent of the pre-COVID level, the weights have been updated using the normal re-weight approach of 2018-19 HFCE data. This represents approximately 80 per cent of the CPI by weight.

For CPI components where the change in expenditure is more sustained and not within five per cent of pre-COVID levels, more up-to-date data sources have been used. This approach represents approximately 20 per cent of the CPI by weight.

The main data sources used to provide up-to-date expenditure information for those CPI components were:

• September 2020 quarter HFCE data;
• Quarterly Business Indicator Survey; and
• Supermarket scanner data.

A full list of data sources for each CPI expenditure class is provided in appendix 1.

The use of HFCE data for CPI weights has many benefits for inflation statistics. The primary benefit is that more up-to-date weights enhance the CPI in its principal purpose as a macro-economic indicator of household inflation. However, there are challenges with using HFCE data for CPI weighting purposes. These challenges were investigated by the ABS and detailed in the information paper Increasing the Frequency of CPI Expenditure Class Weight Updates.

For the 2020 update, movements in the HFCE data for the years 2017-2018 to 2018-2019 were used. The approach for the 2020 update to re-weight the CPI and SLCIs can be summarised as follows:

1. Align the HFCE data with the scope and classifications of the CPI and SLCIs at a detailed product level. This requires the removal of some components of HFCE (e.g. expenditure by Non-Profit Institutions Serving Households (NPISH)).
2. Produce a concordance of the HFCE data to the Consumer Price Index Commodity Classification (CPICC). This provides HFCE data for each CPI expenditure class (EC) for the CPI and SLCIs. These first two steps result in HFCE data aligned to the same concepts and scope of the CPI and SLCIs.
3. Calculate movements from 2017-18 to 2018-19 for the HFCE data from step 2, and apply these movements to each CPI EC to update the expenditure values.
4. Price update the 2018-19 expenditure values for each CPI EC to the September 2020 quarter.
5. Re-scale the price updated expenditure values across all CPI ECs so they sum to 100 (i.e. expenditure shares).

For CPI components that have had a sustained change in spending pattern and   have not returned to within five per cent of their pre-COVID expenditure, more up-to-date data sources were used and additional steps taken:

6. Align the data source with CPI and SLCIs at the lowest possible level and produce a concordance with the CPICC.
7. Calculate a movement from 2017-18 up to the September 2020 quarter.
8. Re-scale the price updated expenditure values across all CPI ECs so they sum to 100 (i.e. expenditure shares). This is done in conjunction with ECs updated with annual HFCE data.

The CPI and SLCI weights reflect the relative expenditures of the CPI population group and SLCI population subgroups as a whole. The weights reflect average expenditure of households and not the expenditure of an 'average household'. The CPI weights for the CPI groups are shown in figure 4.

One thing to note when comparing the weights between 2019 and 2020 is that the weights are relative. The weight of a component of the CPI depends on how expenditure on that component compares to total expenditure (i.e. expenditure shares), rather than the absolute change in expenditure.

For example, if the increase in expenditure for a particular EC is greater than the increase in total expenditure (in percentage terms), the weight for that EC will increase. Conversely, if the increase in expenditure for a particular EC is less than the increase in total expenditure, the weight for that EC will decrease.

This is particularly notable for the 2020 re-weight due to the large fall in the expenditure share for the Recreation and culture group driven by international holiday travel.

For the CPI, households spend the most on Housing (24.05%), followed by Food and non-alcoholic beverages (17.35%) and Transport (10.19%).

Changes in the CPI weights are discussed in more detail below. All analysis refers to the weighted average of the eight capital cities.

Recreation and culture

The Recreation and culture group recorded a significant fall of 4.08 percentage points (pp) to 8.73%, dropping from the third to sixth largest expenditure group in the CPI. The main contributor to this fall was international holiday travel, falling 3.30pp to 0.08%.

International borders remain closed for all holiday travellers in an attempt to reduce the spread of COVID-19. According to ABS Overseas Arrivals and Departures data, overseas arrivals in 2020 were down 99.5% compared to 2019.

Domestic holiday travel and accommodation Expenditure also fell 0.87pp to 2.08%. Domestic travel has made a partial recovery since State and Territory borders re-opened, however interstate travel was still lower than pre-COVID levels reflected in lower levels of domestic airline travel.

Food and non-alcoholic beverages

The weight for the Food and non-alcoholic beverage group rose by 1.60pp to 17.35%.

After lockdown restrictions in March 2020, some social and personal habits changed with restaurant closures forcing consumers to prepare more meals at home. Working from home arrangements remain prevalent, which has maintained elevated levels of grocery spending. This can be seen in  increased spending on food in the Retail Trade data presented in figure 2 of this paper.

To reflect the changes in expenditures for grocery expenditure, weekly supermarket scanner data was used to update the weight.

Alcohol and tobacco

The Alcohol and tobacco group rose 1.20pp to 8.91%. The main contributor to the increase was tobacco, which increased 0.43pp to 3.60%. The continued rise in the tobacco excise tax rate resulted in higher prices and subsequently a higher expenditure share.

Alcohol expenditure rose across all alcohol types, with elevated spending on takeaway alcohol driving the increase. Social distancing restrictions saw a shift in alcohol consumption to takeaway alcohol, which more than offset the fall in alcohol consumed in bars and restaurants. As restrictions eased expenditure on takeaway alcohol remained at elevated levels even while expenditure on drinking in venues returned to pre-COVID levels.

Housing

The Housing group remained the largest component of household spending, with a weight of 24.05%, increasing by 1.12pp.

The New dwelling purchases by owner-occupiers EC remains the largest component of housing costs and was still the most heavily weighted EC in the CPI. This EC increased by 0.65pp to 8.49% as dwelling prices recovered and construction activity remained buoyant. 

Rents was the only component to fall in weight in the Housing group, down by 0.04pp to 6.80%. In recent times there has been low inflation in rents and even falls in rents in some cities as vacancy rates rose and population growth slowed in response to the pandemic.

Electricity rose by 0.17pp to 2.80% as an increase in the number of people working and entertaining at home drove higher usage of residential electricity.

Furnishings, household equipment and services

The weight for the Furnishings, household equipment and services group rose 0.20pp to 8.76%. Most components rose modestly using the 2018-19 HFCE data. Expenditure in these categories remains elevated (see Retail Trade data) but are declining to pre-COVID levels as backorders are completed.

Child care was the main offsetting fall, falling 0.23pp to 0.94%. This fall was driven by price impacts from the return from free childcare offered during 2020.

Transport

The weight for the Transport group fell 0.49pp to 10.19%. The main driver was automotive fuel, which recorded a decrease in weight of 0.41pp to 3.16%. Heavily reduced global demand due to the pandemic, as well as high supply from oil producing nations led to strong falls in the fuel price.

Urban transport also fell 0.21pp to 0.48%, reflecting the strong decline in public transport usage during the peak lockdown period, which was slowly recovering.