Canada’s most prominent House Price Index (HPI) has received a change that could make it unreliable, warns a prominent firm. Oxford Economics is warning clients that the latest changes to the Canadian Real Estate Association (CREA) HPI makes it problematic. The latest change has led to a massive historical revision, minimizing growth over the past few years. As a result, the firm will no longer rely on the HPI, and in some cases won’t be using CREA data at all.
House Price Index
The HPI is the source of the CREA benchmark, or “typical,” home price. Since the composition of sales changes from month-to-month, there can be large variations in average and median sale prices. The HPI, in theory, solves this problem by creating an index. This index makes quantitative and qualitative adjustments, using an approach that’s more typical in appraisal.
To ensure the mix represents current sales, CREA updates the basket annually. Nothing new, but usually they use a process called linking—an adjustment that connects the current benchmark to historical ones. Now they’re dropping linking, resulting in a re-calculation of the index for each year update.
CREA argues this will make the index more reliable, but Oxford Economics is warning clients that the latest modification is more extreme than necessary. It’ll obfuscate trends, making it nearly impossible to follow, which is a benefit and problem, depending on who’s looking at it.
Canadian Real Estate Revision Trims Nearly 40% of Growth
Finding the problems Stillo is referencing isn’t hard, one just has to look at the shift. Starting in 2007 and forward, the HPI has been revised 10% lower. At the same time, the benchmark price increases by 9% at the start of the series in 2005. Historical trends have now been minimized, both showing much smaller corrections.
The gap between the trough and peak are now much closer together—narrowing the gains. In the case of the HPI, price growth from 2005 to 2023 goes from 236% growth to 197% in the newly revised data. That’s one way to trim 39% price growth and make housing “more affordable”, it’s just not the right way.
“CREA’s material revisions to its historical HPI and house prices, nationally and across markets, are problematic for forecasting and make it difficult to maintain a sensible long-term equilibrium house price,” explains Stillo.
Prominent Economics Suggests Using More Reliable Measures, Like The Teranet-National Bank HPI
A central issue here is what the HPI actually measures. Most assume it measures the change in home prices, but that’s not the way Stillo sees it. By using this appraisal-style model, he argues it measures the value people assign to each attribute in a home’s price.
Considering in tight markets, things like location and amenities play a smaller role in determining price, the HPI would be of little benefit to anyone looking to determine trends.
Moving forward, Oxford Economics suggests using more consistent and reliable methods. Their analysis will be placing emphasis on CREA’s seasonally adjusted average sale price. It’s argued that it’s much more volatile than a benchmark, but his pull of the data reveals a smaller variation than the new revision makes.
“While our attention will shift mostly to CREA’s seasonally adjusted average sale price, we will also continue monitoring other less widely reported sources for house prices. High on our list is the Teranet-National Bank HPI which uses a repeat-sales method of price measurement to ensure constant level quality of each property,” says the firm.
Did anyone trust CREA’s HPI? It’s slow and laggy, adjusted and weighed so that it eliminates short-term issues, and maximizes the trend.
Of course if the trend is point in the direction they don’t want, then it needs a revision!
Canada’s real estate data has always been a joke. It’s so expensive to buy property registry data that banks can’t afford it in some provinces.
Obfuscation by hurdles, but most Canadians prefer not knowing how things work. They’re just here to line the pockets of the rich, and hope one day they’ll look at them.
Ranty McRanterson is right in regards to Canada being one of the few countries that doesn’t have any free access to home sale/price data from countries.
In the US the transfer history is public domain and you can just walk into a county office and look up more details for free. Imagine paying $20-30/pop for a whole province, then being restricted from sharing the data because it’s owned by a private company that considers it proprietary? Ludicrous.
There’s no reason that people should not be able to access registry data for free in Canada.
Ludicrous is the nice way of putting it. Corrupt is a more accurate way of describing it.
oooo…. Oxford Econ ahead of the crowd in the call out. I hope this changes how people view the data, and starts some public urgency to release a publicly available data.
I don’t have much faith in the public demanding higher standards though. The loudest voices are in a race to the bottom.
It seems similar to the general trend for Canadian economists to reindex baskets more frequently. Statistics Canada is doing it with CPI, reweighing every year even though they know census data is done periodically because too frequent changes leads to volatile decisions that don’t reflect reality.
Canada is no longer interested in accuracy, but its economists want the smoothest line.
Economics is a strange field, filled with self-important morons that fail to understand real world impact—otherwise they would be in finance.
Doesn’t NS have pretty full, open data on this?
It’s probably, again, a provincial matter, but maybe there is something the fed government can do about housing: legislate transparency on these RE transactions, including who actually owns the property. Having all the information available publicly would be a nice nudge towards some sense of sanity, instead of this information being horded and manipulated by the RE industry, to the detriment of our household finances, our national economy, and our future.
Yeah, yeah, there should be some confidentiality for the purchaser, so maybe require the data be released 90 days post transaction. But beyond that: if you paid and it was foolish, you should look foolish. That just might be all it takes: fear of looking like an idiot.
Canada is full of crooks!
🍒 picked
No one’s sees thre collapse until five years after.
GG NO RE.
I stopped looking at the CREA indexes a while ago. The Teranet National Bank HPIs have their problems but they seem to be much more credible than the CREA indexes. They have also greatly expanded their geographic coverage, pretty much eliminating the only advantage that the CREA indexes had over them.