Yesterday’s inflation data for May got hawks excited at the possibility of another interest rate rise. But we need to start reflecting on exactly what the Australian Bureau of Statistics (ABS) is measuring in the Consumer Price Index (CPI).
“The most significant contributors to the annual rise to May were housing (+5.2%), food and non-alcoholic beverages (+3.3%), transport (+4.9%), and alcohol and tobacco (+6.7%),” the ABS said.
That tobacco is still feeding into CPI despite just 10% of Australians still smoking would strike the British as somewhat peculiar — Britain’s Office for National Statistics’ core inflation measure excludes tobacco, along with energy, food and alcohol. More to the point, the main driver of “inflation” with tobacco is government excise, which is based on… CPI. Transport costs also have a CPI-driven component — state governments lift public transport fares based on inflation (though Queensland’s Labor government has dropped all fares to 50 cents for a while).
As the British example suggests, different countries assess inflation differently. Maybe not better or worse — just differently. And it’s worth remembering if we’re talking about the Reserve Bank making a line-ball call to lift interest rates.
Our primary inflation measures are the CPI and the producer price index (PPI). Both used to be quarterly, but the ABS is expanding its monthly inflation measure — which is currently around 70% of the quarterly CPI — with the aim that, from the end of 2025, it will be a full CPI, with the quarterly index becoming a way of rounding up any extra information or revisions. The CPI only covers eight major cities in Australia — Sydney, Melbourne, Brisbane, Adelaide, Perth, Hobart, Darwin and Canberra, omitting regional areas.
America has three main inflation measures — a CPI and PPI like us, and Personal Consumption Expenditure (PCE) data as well. While America’s CPI covers urban prices movements, the PCE is national. It also includes a broader subset of goods and services prices. Moreover, the weights in the PCE are updated more frequently and better account for consumer substitution. Our CPI uses a fixed basket and can’t track product and services substitution. The US Federal Reserve has targeted inflation in the Personal Consumption Expenditures Price Index (PCEPI) since the 1990s, but only made that index its official inflation target in January 2012.
The US Bureau of Labor Studies produces the CPI while the PCE is produced by the Bureau of Economic Analysts. The former says:
The CPI measures the change in the out-of-pocket expenditures of all urban households and the PCE index measures the change in goods and services consumed by all households, and nonprofit institutions serving households. This conceptual difference means that some items and expenditures in the PCE index are outside the scope of the CPI.
What does this mean for policymakers? In a research paper published in May, St Louis Fed senior economist Christopher J Neely wrote
The CPI and PCEPI both measure the cost of living, but the PCEPI basket of goods and services is broader than that of the CPI … The weights on product categories in the baskets also differ, and the CPI famously weights the cost of housing more than does the PCEPI. Because the PCEPI weights are revised more often than CPI weights, the PCEPI better measures inflation when consumers are changing their buying habits in response to rapidly changing relative prices.”
Neely notes that “CPI and PCEPI inflation are highly correlated, but CPI inflation has exceeded that of the PCEPI by 30 basis points to 50 basis points per year over long periods because of differences in construction.”
That confirms a 2013 observation by the then head of the St Louis Fed, James Bullard, that “CPI tends to show more inflation than the PCE. From January 1995 to May 2013, the average rate of inflation was 2.4% when measured by headline CPI and 2.0% when measured by headline PCE. Hence, after setting both indexes equal to 100 in 1995, the CPI was more than 7% higher than the PCE in May 2013.”
That’s a significant difference, especially if you’re a monetary policymaker weighing up a line-ball rate rise call.
And while our Reserve Bank looks more closely at two artificial CPI measures, the trimmed mean and weighted median, the Federal Reserve doesn’t. Instead, Fed chair Jerome Powell has cited a specific category of inflation — core services other than housing (“a wide range of services from health care and education to haircuts and hospitality … constituting more than half of the core PCE index”) as potentially “the most important category for understanding the future evolution of core inflation.”
Discussion of inflation in Australia, however, is tied tightly to CPI, regardless of how useful a tool that might be for understanding inflation, what’s driving it and where it’s likely to go. The possibility that we could be doing things differently — with different consequences — doesn’t get much attention.
