Canada’s real estate market has skyrocketed over the past year, setting off a vigorous discourse over the dangers of rapidly rising prices and steeper barriers to homeownership.
And yet, inflation numbers offer a subdued – even drab – version of events.
As bidding wars and six-figure price increases become the norm, it’s tough to find much of an impact on Statistics Canada’s go-to measure of inflation, the consumer price index.
The price of shelter rose 2.4 per cent in March from a year earlier, as did owned accommodation. That was in line with overall inflation, which grew 2.2 per cent. Meanwhile, the average selling price of a home surged by a whopping 32 per cent over the same period, according to the Canadian Real Estate Association.
How to douse Canada’s housing mania when you can’t raise interest rates
The disconnect is just the latest chapter in a long-running debate over how to treat housing in inflation statistics, with no consensus emerging in advanced economies. After all, homes are a major expense for the typical family, but they’re also an asset – unlike, say, a television purchase.
Still, the current situation raises questions about whether CPI is sending mixed signals to the Bank of Canada, which uses 2-per-cent CPI inflation as its monetary-policy target, and whether it’s damaging the credibility of an indicator that’s already eyed with suspicion by those accustomed to sticker shock.
“The average Canadian didn’t need a PhD in economics to be able to tell that something was amiss,” said David Williams, vice-president of policy at the Business Council of British Columbia. “CPI would tell you that the shelter-cost inflation was benign in Canada, and I don’t think that really lined up with people’s lived experience.”
To a large extent, the discrepancy is by design. Statscan doesn’t intend to track short-term changes in housing prices; it’s trying to assess the continuing cost of housing, not the cost of houses.
Inflation is easy to measure for rental properties, where Statscan can track changes in monthly rent. It’s more complex for people who occupy their own home. Statscan’s solution is to treat homeowners as if they’re renting to themselves, and includes mortgage payments, alongside other expenses a landlord would typically face, in determining inflation. Notably, down payments are excluded.
“CPI does not include the purchase of a property, because in this case, we don’t consider a house a consumer good,” said Heidi Ertl, director of the consumer-prices division at Statscan. “We consider it an asset.”
One contentious element of CPI is the Mortgage Interest Cost Index (MICI), which is one of the main places where house prices show up in CPI. The index tracks mortgage costs for homeowners by looking at changing interest rates and house prices over time. (It uses a weighted average approach, that includes historical data but puts more weight on recent data).
The problem is until this year, Statscan relied solely on sales data for newly constructed homes when calculating the MICI.
New homes were “basically a proxy for all housing prices,” Ms. Ertl said, adding Statscan didn’t have a “robust and reliable” source of data for the resale market.
This approach tended to underestimate the size of mortgages Canadians were taking on, Mr. Williams said.
“Established house prices are up 200 per cent over the last 15 years, while newly constructed home prices were only about 20 per cent. These are orders of magnitude different,” he said of the Vancouver real estate market.
Statscan recently tweaked its methodology, adding resale data from six major metropolitan areas in February. The change is “important for accuracy of the CPI,” Ms. Ertl said.
A second issue, which the inclusion of resale data does not address, is the impact that interest-rate changes have on the Mortgage Interest Cost Index.
When interest rates drop, the index declines, as homeowners take on new mortgages or refinance existing mortgages at lower rates. Over the past year, the MICI declined 6.3 per cent, owing to Bank of Canada rate cuts and its continuing efforts to hold down interest rates through its quantitative-easing program.
At the same time, however, the central bank’s ultraeasy monetary policy is throwing jet fuel on the housing market. This creates the uncomfortable impression that the bank’s main gauge of inflation and guide to monetary policy is unresponsive to the segment of the economy that the bank is stimulating the most.
“When interest rates go down, other things being equal, housing prices and inflation pressures are likely going up, while the standard CPI method would report them as going down,” said Finn Poschmann, senior fellow with the Fraser Institute.
“That leads to a cost-of-living indicator that doesn’t quite reflect what consumers see and feel, and an inflation indicator that doesn’t quite reflect the long-term cost of owned housing relative to other things we buy,” he said.
To that point, there is a consistent mismatch between CPI inflation as Statscan measures it and how Canadians typically perceive inflation. In an online survey conducted by the Bank of Canada last year, 55 per cent of respondents said 2-per-cent inflation was not a “realistic representation” of their experience of inflation, while 66 per cent of respondents believed inflation in Canada is generally higher than 2 per cent.
Some countries get around the problem of interest rates by excluding owner-occupied shelter from their cost-of-living indexes altogether. Others, such as New Zealand and Australia, focus solely on home-acquisition costs when calculating inflation for owned accommodation.
This latter approach does a better job tracking inflation in the real estate market. But it’s less well suited to tracking continuing shelter costs, given only a small portion of homeowners buy houses each month.
The bank’s own researchers, in 2015, advocated replacing the mortgage-interest-cost component of CPI with a “financial-opportunity-cost” metric, which would do more to account for the cost of owner’s equity, including down payments, and house-price appreciation.
Still, the bank welcomed the inclusion of resale data into the MICI, “as we do all improvements to CPI that more accurately reflect how Canadians experience inflation,” bank spokesperson Sean Gordon said.
It will take time for that change to materialize in the numbers. When calculating the MICI, Statscan includes historical data on both mortgage sizes and interest rates – the idea being that only a small portion of Canadian homeowners are buying properties each month, so most homeowners face historical costs on a continuing basis.
“What you might see is a little bit of a hockey-stick effect. Meaning that by the time [the new resale data] shows up – because again it’s going to be averaged over 25 years – interest rates might be higher because the economy has normalized and the Bank of Canada can take its foot off the accelerator,” Mr. Williams said.
“So you have higher interest rates acting on a higher principal amount because you’ve made this measurement change, that could move the mortgage-interest component of the CPI substantially higher,” he said.
In the meantime, cost pressures are mounting in other parts of the CPI related to shelter, namely the homeowners’ replacement cost index. This is essentially a measure of depreciation, reflecting the hypothetical cost of replacing the lost value from a stock of dwellings.
To determine that cost, Statscan looks at the price of new home structures. And new-home prices have soared over the past year, owing to substantial demand and the rapidly climbing price of lumber.
Your time is valuable. Have the Top Business Headlines newsletter conveniently delivered to your inbox in the morning or evening. Sign up today.