The Bank of Canada is considering ending the practice of calling certain core inflation gauges its “preferred” measures, a shift that would change how investors interpret monthly data.
In a speech at Western University’s Ivey Business School on October 2, Deputy Governor Rhys Mendes said the bank is reassessing both the tools it uses to extract the inflation signal from noisy data and the way it talks about those tools.
“Should we broaden our list of preferred measures? Or perhaps even end the practice of identifying some measures as ‘preferred’?” Mendes said. He added that “our target is overall inflation, not core inflation.”
Mendes sketched a landscape where no single indicator can be trusted to tell the whole story.
The bank has long relied on CPI-trim and CPI-median to filter out volatile price moves after discontinuing CPI-common as a preferred gauge in 2022 due to revisions.
Those measures “have generally been helpful,” he said, but at times can mislead.
In August, headline CPI ran at 1.9 percent while the bank’s preferred core measures were about 3 percent and several alternative core measures clustered closer to 2.5 percent.
The Governing Council’s assessment was that underlying inflation sat “in the vicinity of 2.5 percent,” he said.
What could change
One fix under review is how to treat mortgage interest costs, which mechanically rise when policy rates go up. Mendes noted that this component can obscure the broader response of prices to monetary policy.
He pointed out that CPI-trim did start filtering out mortgage interest costs after rate hikes, but because the category is large, “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 percent, but an alternative version that pre-excluded mortgage interest costs was 2.1 percent. The bank is weighing whether to build that pre-exclusion into all its core gauges.
The central bank is also testing fresh approaches, including a multivariate core trend inflation model that aims to isolate the persistent part of price growth and separate common economy-wide pressures from sector-specific ones.
Researchers are exploring whether clustering algorithms can improve core measurement as well. Early results look promising, though the new methods can be sensitive to revisions, Mendes said.
The rethink arrives as the Bank of Canada and the federal government gear up for their regular renewal of the monetary policy framework in 2026.
Mendes said the five-year review is a chance to refine how the bank assesses underlying inflation and communicates that assessment in a world more prone to large shocks.
He emphasized that while core metrics remain an important input, policy decisions rest on a broader suite of indicators, not a single reading.
Traders have often keyed off CPI-trim and CPI-median to handicap rate moves.
If the bank formally broadens the set of measures it highlights, monthly reactions could hinge less on two numbers and more on a wider dashboard that includes breadth gauges across the CPI basket and estimates of underlying trend inflation.
That could smooth knee-jerk swings around individual data points and re-focus attention on the bank’s 2 percent target, which remains the anchor for policy. Mendes’ message boils down to this: there is no silver bullet for parsing inflation.
