Consumer spending categories (more commonly known as “baskets”) are used in measuring consumer price inflation. Generally, these spending baskets remain relatively stable over the short-term. As a result, statisticians make infrequent update to these baskets (in Australia spending basket are updated annually using consumer spending data from 18 months beforehand). Consumer spending patterns can change over the long-term from “substitution bias” which occurs when price increases cause consumers to substitute towards alternative (cheaper) products. However, economic shocks (like the current COVID-19 pandemic) also produce changes in consumer spending patterns.

COVID-19 has caused a substantial shift in the demand for goods and services as industries have been closed and mobility has been restricted. Generally, demand for goods has increased while services demand has fallen. Measured inflation across developed economies has been very low over the first six months of 2020. Australia even had deflation with the headline consumer price index (CPI) decreasing by 1.9% over the June quarter. However, these inflation measures are using pre-COVID consumer spending baskets. Prices for high-demand goods have risen while low demand services prices have fallen. If inflation data was updated for COVID-19 spending patterns, measured inflation would not look as low as it currently appears. We go through these issues in this Econosights.

Inflation in Australia in the time of coronavirus

COVID-19 has led to falls in services prices (due to the temporary shutdown in some service sectors) but stronger growth in goods prices (see chart below), with prices even going up for some goods related to the home (things like household equipment to help with working from home requirements, perishable food, cleaning products and toilet paper).

Temporary supply disruptions amid the start of the pandemic in March-April also led to higher prices for some goods that were in high demand. The chart below shows the divergences in prices for items that have been high in demand during the pandemic versus others that have been low in demand.

Along with wide divergences in price changes, Australian consumer spending patterns have shifted more significantly over the past six months compared to the usual annual updates done for spending baskets. The biggest pandemic-related increases in spending have occurred in food and alcohol, furnishings and housing while the largest falls in spending have been in eating out, transport and recreation and culture. These areas that have had the greatest rises and falls in spending have also seem some of the largest price increases and decreases in the inflation data.

If we re-construct the Australian June quarter CPI using our estimate of what the consumer spending baskets actually were during the COVID pandemic (using consumer spending data) then inflation would look quite different. The headline CPI would show a fall of only 0.1% over the June quarter, rather than the reported fall of 1.9% which still shows very low inflation, but not quite the significant deflation picture painted by the headline figures. This notable difference in actual versus estimated inflation reflects that spending has lifted on items where prices have also increased. If we repeated this exercise in other developed countries, like the US, we would expect a similar conclusion.

Consumer inflation expectations

The idea for this Econosights came about because I was often told that the CPI data wasn’t an accurate reflection of reality during the pandemic which led me to think about consumer inflation expectations and the change in consumer spending patterns over 2020. Historically, consumers tend to overestimate actual inflation outcomes (see the chart below).

In “normal” times, the divergence between consumer inflation expectations and actual inflation could be because prices of essentials like healthcare, utilities, fresh food and vegetables and education tend to have solid above-average price rises over time. But, discretionary items that consumers want to spend money on (like clothing, technology based goods and communication services) have generally had price falls over recent years. Consumers tend to feel the pain of paying for items they see as essential and remember the associated costs.

A recent speech from the Bank of Canada Deputy Governor highlighted some of the reasons why consumer inflation expectations continually tend to be higher than actual inflation outcomes. These include:

• Consumers may feel that their spending baskets are different to those used in the overall CPI. The Bank of Canada and Statistics Canada constructed varied consumption baskets for different households to test this theory but still found that the average inflation rate tracked close to the overall headline CPI.
• Consumers may not be accounting for the improvement in quality that occurs with higher prices (meanwhile statisticians account for quality improvements in the CPI). Quality improvements in goods items are often taken for granted because it’s hard to measure.
• Consumers may be including home prices as part of their spending basket. House prices are not included in CPI, because they are considered an investment, rather than a consumption good (like other parts of the CPI data). The cost of living (rent, household services and costs associated with new construction) are included in the CPI, but price growth in these categories has been well below home prices. Home price growth has been extraordinarily higher than inflation over time (see chart below for the Australian experience) although with more fluctuations.
  

  ---

Implications

Headline CPI measures during the COVID-19 pandemic appear to be understating actual consumer inflation. Consumer spending baskets are likely to remain different to their pre-COVID norms for a while. Future economic shocks could also see similar changes to spending behaviours (more spending on goods and less on services). Understating consumer inflation could mean that cost of living burdens is being underestimated for consumers. For example, the big drop in headline CPI could mean that real wages growth (nominal less inflation) shows up as a big real wage increase (see chart below). But this is an overstatement. As well, many parts of the economy are tied to CPI outcomes, including government welfare payments and business wage outcomes.

Statisticians making more frequent updates to consumer spending baskets in CPI data may not be feasible but perhaps there is a case to make these updates during big economic shocks, like the current pandemic to help understand cost of living issues facing consumers. And these issues emphasise the need to use the core measures of inflation more widely as these measures are less subject to extreme distortions in quarterly price movements.