Canadian mortgages are remarkably low-risk assets, at least according to historical data.
BMO Capital Markets put out a recent report which underlined that point. Since 1979, loan losses on uninsured mortgages have averaged a paltry 2-3 basis points, notes author John Reucassel. That’s just $20-30 per $100,000 of mortgages.
Even during the 1990 recession, uninsured mortgage losses topped out at 6 bps. (The peak was in the early 80s when credit losses hit 12 basis points.)
You’d expect insured mortgages to be more risky, but from a loss standpoint, CMHC’s losses have averaged a very reasonable 9 bps over time. In the big scheme of things, 9 bps is small potatoes. Remember that CMHC charges borrowers up to 275 basis points (of their principal) on a typical insured mortgage.
*********
Despite the above, the past is not the future. Canadians owe more money than ever and average home prices are stratospheric in some areas. So what can we expect for future loan losses?
“We believe the credit risks will be more closely related to unemployment or a rapid rise in interest rates (at least 300 bps) than a house price correction,” says BMO.
Statistically, however, the main triggers of mortgage defaults are “the three “Ds” – death, divorce, and dismissal (i.e., job loss).” And, of those three, only the last one is a systemic threat.
Regardless of what deflates the housing balloon, the odds are high that nationwide uninsured loan losses will eventually top the 12 basis point mark we saw in the 80’s. How much so is anyone’s guess. That’s why it’s fortunate that banks and insurers can withstand multiples of that level.
Rob McLister, CMT
Last modified: April 25, 2014
Hi Rob,
Do you know of anyone who publishes loss statistics by loan-to-value?
Unfortunately not. But insurance premium tables provide a sense for how insurers view loss risk by LTV:
<= 65% → Premium: 0.50% <= 75% → Premium: 0.65% <= 80% → Premium: 1.00% <= 85% → Premium: 1.75% <= 90% → Premium: 2.00% <= 95% → Premium: 2.75% Source: CMHC
I missed how you got from loan losses never exceeding 12bps in any one of the significant down markets in the last 30+ years, to the odds being high that losses will be higher going forward. I think you convinced me of the opposite with your well written article.
Hi BA,
Thanks for the note. I’ll answer a question with a question. :) What are the odds that most macro-economic records will be broken? One could argue that they’re well above 50%, given a long enough time horizon.
In the case of housing, we’ve never had valuation metrics like we see today. While I’m confident that underwriting is as good as it’s been in years, if these metrics remain stretched, a macro shock (e.g., severe unemployment, an unexpected 300+ bps rate spike, etc.) could–in my view–potentially trigger losses exceeding 12 bps. I’m not saying it will happen tomorrow, but I’m comfortable in projecting that it will happen eventually.
Unfortunately that is a government assessment of the premium necessary. I would love to see what private industry would charge for mortgage insurance without government backing…