The Bank of Canada is thinking about stopping the practice of calling some core inflation gauges its “preferred” measures. This would change how investors look at monthly data.
On October 2, Deputy Governor Rhys Mendes spoke at Western University’s Ivey Business School and said that the bank is looking at both the tools it uses to get the inflation signal from noisy data and how it talks about those tools.
“Should we add more measures to our list of favourites? Or maybe even stop calling some measures “preferred”? Mendes said. He also said, “Our goal is overall inflation, not core inflation.”
Mendes drew a picture of a landscape where no one sign can be trusted to tell the whole story.
The bank has used CPI-trim and CPI-median for a long time to get rid of price changes that are too big. They stopped using CPI-common as a preferred measure in 2022 because it was changed.
More Read
He said that those measures “have generally been helpful,” but they can also be misleading at times.
The CPI was 1.9% in August, but the bank’s preferred core measures were about 3%, and a few other core measures were closer to 2.5%.
He said that the Governing Council thought that underlying inflation was “in the vicinity of 2.5%.”
What might change
One possible fix is to look at how to handle mortgage interest costs, which automatically go up when policy rates go up. Mendes said that this part can hide the bigger picture of how prices react to changes in monetary policy.
He said that CPI-trim did start to leave out mortgage interest costs after rates went up, but because the category is so big, “its persistent exclusion limited CPI-trim’s scope to exclude other, more temporary, sources of upward pressure on inflation.”
In December 2024, CPI-trim was 2.5%. However, a different version that didn’t include mortgage interest costs was only 2.1%. The bank is thinking about whether or not to add that pre-exclusion to all of its main measures.
The central bank is also trying out new ideas, like a multivariate core trend inflation model that tries to separate the long-term part of price growth from the short-term pressures that affect the whole economy and those that only affect certain sectors.
Researchers are also looking into whether clustering algorithms can help with core measurement. Mendes said that the new methods can be sensitive to changes, but the early results look good.
The rethink comes at a time when the Bank of Canada and the federal government are getting ready to update the monetary policy framework in 2026.
Mendes said that the five-year review is a chance to improve how the bank looks at underlying inflation and how it talks about that in a world that is more likely to have big shocks.
He stressed that core metrics are still an important piece of information, but policy decisions are based on a wider range of indicators, not just one reading.
Traders have often used CPI-trim and CPI-median to help them figure out how to bet on rate changes.
If the bank officially expands the list of measures it focuses on, monthly reactions might depend less on two numbers and more on a larger dashboard that shows how the CPI basket is doing as a whole and how inflation is likely to change over time.
This could help stop knee-jerk reactions to individual data points and bring the focus back to the bank’s 2% target, which is still the main goal of policy. Mendes’ main point is that there is no simple way to figure out inflation.
