I’ll confess I’m a little gun-shy when it comes to making dire predictions about Canada.
When you write about global markets for public consumption on a daily basis, you make an idiot out of yourself in the eyes of someone, somewhere every single day. That’s just part of it.
After all, for every market you opine on, there’s an expert out there somewhere. And by virtue of the fact that you can’t, yourself, be an expert on every single market, pretty much every post you publish is going to seem amateurish to someone.
And that’s fine. Because you’re never going to talk to, let alone meet, the vast majority of those experts, so you know, “fuck ’em.”
But when it comes to Canada I actually converse on a fairly regular basis with two people who know more about it than I do – namely Bloomberg’s Luke Kawa and the excellent Kevin Muir, who writes The Macro Tourist. That puts me in the uncomfortable spot of having to possibly answer for anything stupid I say about Canada.
Hence the gun-shy bit.
After all, how are you going to pretend like you know more about Canada than a guy who wears these socks:
I don’t know if Kevin has any knee-highs like that.
That said, what I do know something about is what happened in Canada just prior to the crisis. Indeed, there was a meltdown in the Canadian asset-backed commercial paper market that predated the meltdown in the US and it was triggered (if unjustifiably so) by the early signs of a collapsing US housing market.
I won’t go into the details here, but I’d wager I know more about that episode of Canadian financial history than almost anyone who wasn’t directly involved.
So when I read stories about Canada’s housing bubble, I’m always tempted to weigh in – even if it means getting something woefully wrong and having Luke tell me that I’m just as uninformed as everyone else about his beloved country.
This whole conversation took on a bit of extra urgency of late after Stephen Poloz was effectively forced to “take one for the team” (as it were) when it came to “proving” that central banks were serious about actually pulling the trigger on normalization.
I imagine the conversation went something like this:
Yellen, Draghi, and co.: “Look Stephen, you’re going to have to hike rates. We know you’ve got a housing bubble and yes, we know the bottom could fall out for oil prices at a moment’s notice, but you’re the only one who can hike without causing a global panic in risk assets, because let’s face it, no one gives a shit what you guys do.”
“But see if you hike, then we can at least point to the BoC as evidence that we, as a group of central bankers, are trying. And besides, wouldn’t you like to burn all those loonie shorts?”
“So what we’ll do is send Yellen out dovish on Capitol Hill and that will drive yields down and boost risk assets and you can hike at the same time (literally) and no one will even notice.”
On Wednesday, I get to have my cake and eat it too. Because Goldman is out with a new note called “Does Canada 2017 = US 2007?”
So what I get to do here is present excerpts from that note for you to consider and if there’s anything flawed about the analysis, well then I can just blame Goldman.
Focus: Does Canada 2017 = US 2007?
Nominal home prices in Canada have grown by 13% over the past year, and by 200% since 2000. These sharp increases in home prices in Canada have invited comparisons with the US housing market in the period leading up to the Global Financial Crisis (Exhibit 1). A large downturn in the Canadian housing market in 2017 would seem particularly untimely given the likelihood of rising mortgage rates in Canada. Most Canadian mortgages have 25-year amortization schedules but 5-year terms, and so borrowers typically have to re-qualify for new mortgages every 5 years. Under our rates view, many Canadian mortgage borrowers may be forced in the coming years to refinance their loans at higher mortgage rates. Given the potential risks that a housing downturn could pose to the Canadian economy, we address here the questions (1) “is Canada’s housing market in 2017 comparable to the US’s in 2007?”, and (2) “will rising interest rates lead to significant payment shocks and mortgage defaults among Canadian homeowners?”.
With respect to the first question, our answer is “in some respects, Canada in 2017 and US in 2007 are similar, but in many respects they are not”. One important difference is with respect to the mortgage lending standards prevailing in the two times and places. Exhibit 2 charts an indicator of US lending standards vs. the US house price index, showing that standards were still loosening during 2005-2007 even as house prices were approaching a peak. In 2006, over 40% of US mortgages were funded via the non-agency RMBS market, where no-doc and low-doc lending and usage of “exotic” loan products such as negatively amortizing adjustable rate mortgages were most common. By comparison, Canadian banking regulators have generally tightened lending standards since the financial crisis, including reducing maximum LTV ratios and amortization terms, and, more recently, Vancouver and Toronto have introduced foreign buyer taxes to dampen house price growth. Some non-bank lenders in Canada have been introducing riskier “Alt-A” mortgage products in recent years, but these are less prevalent than were the comparable products in the US before the housing bust.
Indeed, Canada’s mortgage delinquency rates have remained structurally lower than the delinquency rates in the US for the past several decades (Exhibit 3). These lower delinquency rates reflect a combination of more conservative underwriting and more lender-friendly mortgage foreclosure laws in Canada. The lower baseline mortgage default rates in Canada, combined with stronger bank capitalization, suggest that a house price decline in Canada of the magnitude experienced in the US during 2007-2011 would likely pose smaller systemic risks than were realised in the US during the financial crisis period.
Exhibit 3 shows that delinquency rates in the US had already started trending up by 2007, whereas delinquency rates in Canada remain low, suggesting that Canada’s mortgage market is at least not yet at the same distressed stage as the US’s was in 2007. Similarly, Exhibit 4 shows that the mortgage debt service ratio in US was historically elevated by 2007, whereas the comparable ratio in Canada remains near historically normal levels. Rental vacancy rates and unsold housing inventories are among the other indicators that were already showing stress in the US by 2007, but which still appear healthy in Canada as of 2017.
Whereas backward-looking indicators of the Canada housing market may still appear relatively benign, market observers have expressed concern that rising interest rates will lead to significant stress, since most Canadian mortgages have only 5-year terms and thus will need to be refinanced at higher rates. Our calculations suggest that this risk, while non-negligible, may not be disastrous. As an example, a borrower who took out a 25-year amortization, 5-year term, 4% mortgage for $300K in 2013 would have been paying a monthly principal plus interest payment of $1,584. If this borrower is forced to refinance the remaining $262K balance in 2018 at a higher 6% rate – a fairly extreme scenario – and takes another 25-year amortization loan, the monthly payment would increase to $1,688, just a 7% increase relative to the original loan payment. The payment increase is relatively modest because (a) the rising interest rate leads to a higher interest payment but also to a smaller monthly principal payment; and (b) the original 2013 loan was scheduled to fully amortize by 2038, but the new 2018 mortgage is not scheduled to pay down until 2043. By extending the final amortization date, the payment shock is partly mitigated. We thus do not expect a large pickup in mortgage defaults due to higher payment burdens among existing borrowers.
While rising rates may not induce a surge of defaults among existing borrowers, higher rates can negatively affect housing affordability for future potential homebuyers. Our previous research in the US and other international housing markets has suggested that if higher rates are associated with a strengthening economy or stronger inflation, then the overall impact on house prices can be limited, but if higher rates are driven by an adverse policy shock, the impact may be more negative. In the case of Canada, we expect rising rates to be correlated with a stronger labor market, and thus expect the negative impact on housing affordability to be real but manageable.
In summary, the recent rapid rise in house prices in Canada presents a risk of eventual over-heating. A model of bust risk that accounts for house price-to-rent ratios, past changes in real house prices, investment-to-GDP ratios, real GDP growth and inflation puts the probability of a 5% or larger downturn in real house prices over the next 5-8 quarters at around 30%. At the same time, we see significant differences between the US housing market in 2007 and the Canadian market in 2017, and for this reason we think it may be early to look for a downturn in Canadian house prices of close to the magnitude seen in the US before the financial crisis.