What does an insured mortgage borrower look like?
According to CMHC data from last quarter, residential borrowers who paid for mortgage default insurance had:
-
8% equity on average (92% loan-to-value)
- Less than 10% down in almost 7 out of 10 cases
- 25-year amortizations, on average
- Amortizations over 25 years, only 3.1% of the time
-
745 average credit scores
- 82.6% had scores of 700+ (scores range from 300 to 900)
- CMHC doesn’t have a 680+ breakpoint in its credit score data but hopefully it adds one. A score of 680 is the minimum needed to qualify for higher debt ratio flexibility, as well as the best interest rates in many cases.
- Average mortgage amounts of $230,416
-
Mortgages over $600,000, only 4.7% of the time
- About the same as last year
-
25.5% average gross debt service (GDS) ratios
- This is quite low as 32% is the traditional maximum
If we look at all transactionally-insured borrowers on CMHC’s books, not just those who closed last quarter:
-
9.3% have less than 10% equity
- This speaks to both surging property values and people paying down their mortgages
-
51.4% took out amortizations over 25 years
- This ratio will steadily drops thanks to the extinction of high-ratio amortizations beyond 25 years
-
730 is the average credit score
- Recent borrowers have higher average scores than CMHC’s overall book.
CMHC should be commended for expanding the borrower data it discloses publicly. It shares a lot more than the other two insurers and the data thus far shows no obvious areas for concern.
That said, it still doesn’t provide some of the most important risk metrics of all. Most obvious is the lack of data on:
-
Total debt service (TDS) ratios
- How close is the average government-backed borrower to the traditional 40% maximum debt ratio?
-
TDS by loan-to-value
- How many people have a high TDS and less than 10% down?
- This is one of the most telling metrics of all because folks with low equity and high debt ratios are among the riskiest prime borrowers
-
Average loan amount by loan-to-value
- How big are the mortgages for people with only 5% down?
CMHC could answer all of these questions for the Canadian taxpayers who back the crown corporation, and hopefully it eventually does.
Mind you, if this data were available, it’s unlikely that it would reveal any serious cracks in CMHC’s hull. But we won’t know that until we see it.
Rob McLister, CMT
Last modified: September 2, 2014
If my suspicions are correct, I would expect to see those people in the acquisition phase of their lives (ie 25-45 years and the target demographic for the mortgage industry) to have the highest TDS while dealing with the lowest equity in their home. I’m confident many older families remember the days they were there, struggling to keep the roof over their heads; why should the younger ones have it any different?
Today and yesteryear are not a fair comparison. Most older families didn’t buy $400,000 houses at 9 times income with 5% down, 8 year car loans and $25,000 of credit card debt.
More disclosure is a good thing, however I did chuckle a little when you suggested taking TDS seriously.
I don’t mean to say it isn’t an important metric in the underwriting process, but in my experience it has traditionally been misunderstood. Whenever I’ve seen “big data” on portfolio-level TDS distributions, there’s always a huge spike right around the TDS threshold (eg. 40%). And yet, in reality the majority of people aren’t actually living on the edge.
This discrepancy between data and reality is explained simply enough: If a file already has good enough TDS, there’s often no incentive to continue polishing the file and lower TDS further. Why waste time proving additional income, getting liability payments adjusted, etc when you could be moving on to your next deal?
Don’t let the continuous nature of the variable fool you, in practice TDS is barely more than a binary pass/fail metric.
Hi Banker,
Thanks for the note. The question isn’t so much whether the majority are living on the edge. We all know they’re not. The question is really: how much of the minority is vulnerable to a shock?
You make an excellent point that partly explains why TDS, in isolation, is not a perfect barometer.
On the other hand, the trend in distribution of borrowers with high TDS and low equity could be quite telling indeed–especially when drilled into.
Unfortunately, we don’t have that data because CMHC doesn’t release it. Hence this story.
Averages cover up a lot of problems but if CMHC touts average GDS they should disclose average TDS. They go hand in hand.