Default Criteria & Amortizations

DefaultsFitch Ratings, one of the big three North American credit rating agencies, published its mortgage loss model today. It’s basically a bunch of formulas and assumptions that can be used for estimating losses on prime mortgages.

Fitch researched hundreds of thousands of mortgages and settled on six primary factors that drive defaults. They are (in order of significance):

  1. Borrower equity (the strongest driver of defaults)
  2. Credit score
  3. Total debt service ratio
  4. Loan purpose (purchase or refinance?)
  5. Occupancy (owner occupied or investor?)
  6. Property type (single family or multi-family)

As you may notice, amortization length was not deemed a significant factor that influences arrears. In fact, it is pretty far down the list of items causing delinquency. Yet, as we saw this week, longer amortizations still get a bad rap in regulator circles.

But how much do extended ams increase systemic risk? Is it possible that their flexibility actually reduces risk for many borrowers? This week’s column from the Globe and Mail touches on that: Would shorter amortizations make the housing market safer?

Rob McLister, CMT

  1. The Fed’s goal is not to make the market safer, they want to make mortgages harder to get and take the “air out of the bubble”. But I guess that in turn does make the market safer. Increasing the minimum down-payment or credit score or lowering TDS/GDS ratios would have the same effect of making mortgages harder to get.

  2. As a private lender, I find that when people lose their jobs, and there is lots of equity, somehow they managed to make their payments. Conversely, when people have no equity or minimum, with or without jobs, when things get tight, they ‘no longer’ can “afford” the house, and default occurs more frequently. In fact, with negative equity, this can be the actual cause of default, with some homeowners who are well equipped, income wise, to handle the payments choosing default as a financial planning strategy. So I am partial to the quoted number #1 default as a cause whether indirect or sometimes direct.

  3. Hi Jim, The Finance Department has more than one goal but yes, slowing the market is a priority. This post was primarily meant to address the misconception that extended amortizations are unduly risky. Cheers…

  4. The latest Fitch report which reviewed hundreds of thousands of mortgage to determine the mortgage default factors is another study which confirms that reducing amortization periods has little to do with reducing mortgage default – but has strangled real estate activity [ sales VOLUMES ] – Federal Dept of Finance seems quite smug about the slow down – it has little effect though on reducing ” household debt ” – their main argument these past few months. In fact it will probably increase household debt as Canadians seek more expensive debt payments.Perhaps the Federal Finance dept. should be more focused on credit card debt and Federal Debt instead of stopping growth.

  5. I think that longer amortization lengths would not directly contribute to mortgage defaults. The problem is when you allow longer amortization with an increased amount of dept (more than 3X your income) at the lowest interest rates in history. If borrowers were indefinitely limited to 3X the household income, longer amortization would simply reduce and extend you mortgage payments while increasing the amount of interest paid to the bank.

  6. This information is something many of us have seen before but it is simply so true.
    I remember seeing the stats out of the mortgage insurers showing that the default rate from deals with Beacon Scores over 800 were effectively zero.
    As a private lender the equity factor is such a huge driver: a 60% LTV deal always finds a solution to a problem even if it is just refinancing away from me and the problem becomes someone else’s.
    This list of underwriting eternal truths trumps so much of the analysis and commentary we see today.

  7. from the article.. “Fitch currently estimates that home prices are overvalued by approximately 20% in real terms across Canada (with regional variations).” … “Two-step rating stress. House prices are first reduced to their sustainable value and then subjected to a further stress that corresponds to each rating category. The ‘AAA’ market value decline stress assumes that home prices decline below the sustainable value by 35%.”
    so given current prices, anyone who’s mortgage does not have atleast a 55% equity position would not meet AAA in their new ratings system. Will this result in a lower rating and higher borrowing costs for non insured mortgages with 25%-35% down going forward?

  8. In my view, who cares what amortization people get as long as they are no risk to anyone else. It’s time for people to stop trying to babysit other responsible adults.

  9. Obviously, longer amortizations can extend the repayment period and increase the total interest. Tell us something we don’t know.
    What are you really saying here? That you know best when it comes to how much interest people should pay? Should people call you first to discuss their budget before buying a house? Why don’t we all just buy studio condos so we don’t have big mortgages. That would save way more interest than a 25 year amortization. Three kids in a 500 sq ft condo may be cozy, but think of all the savings.

  10. Initial amortization and final amortization are two different things. Someone with a 35 year amortization can pay off their mortgage faster than someone with a 25 year am. It just depends on their prepayments.

  11. Responsible ones don’t need supervision, I agree.
    I am amazed about people that have increasing debt after years of paying thousands in interest every year.

  12. Full agreement with Ron, equity & proven income trump just about any other “analysis to paralysis” parameters FI s want to use. Speaking to some fairly senior people at CMHC a few years back, i was told between two deals , one with 5% down , another with 10% down, and all other things being equal, the 5% down deal had a 40% greater chance of default. Again, nothing to do with amortization.

  13. The amortization length isn’t on the list as we have only had the lowered ams for about a year. Typically, you will have to see exposure to this type of chance for a couple of year to observe the results. Also, think of the fact that most people who took a 5 year term and a 40 year am in 2008 were at about 5.75%. Now they are renewing into a 5 year 2.84% rate on a 25 year amortization and lowering their payments. So the amortization change likely will NOT affect them. The issue is that folks today may not be able to refi from 3.89% to 2.84%, drop from 30 years to 25 years and save money too.

  14. Someone’s a little sensitive!! What I’m saying is that I don’t think there’s anything wrong with increasing the amortization period if the amount of dept is restricted to approximately 3 times the household income. If you can’t afford to buy a house don’t buy one.

  15. >> Now they are renewing into a 5 year 2.84% rate on a 25 year amortization and lowering their payments.
    No. They are grandfathered in. As long as they aren’t changing their mortgage they would be on a 35 year term.

  16. “This post was primarily meant to address the misconception that extended amortizations are unduly risky” -Rob McLister
    Of course, extended amortizations are risky, precisely because they fuel the kind of irrational price increments that ultimately end in a bubble.
    At 25-year amortization, a buyer earning 50K per annum, to take a totally random example, can only pay so much for a house. And when buyers can only afford so much, the seller, ultimately, will only be able to sell for so much.
    However, if you now now extend the amortization from 25 to 40 years, or 60 years, or to 100 years, everything suddenly changes! Your average buyer can now pay much more for a house, which means that the average seller will also be able to sell for so much more, which leads to the kind of ridiculous bubble prices that the Feds are trying to tone down.
    So if you are trying to put the brakes on rising house prices, amortization a key tool -about as important as the interest rate.
    A real estate agents earns better commisions if the house sells for 1 Million, instead of selling for 300K. A mortgage broker earns a bigger commission if the mortgage is 1 million, instead of 300K. So
    of course, when selling prices are huge, agents and brokers earn jumbo commissions, which is really the only reason why these people are so strongly invested in this inflated status quo.
    The current possible amortization on non-insured mortgages is a serious loophole on what the Feds are trying to accomplish. If he is serious about getting a handle on the stewing housing bubble, Flaherty should promptly close that loophole, no questions asked. 25 years is a lot of time to pay for any item. If you can’t afford a house with a 25 year amortization, you should simply not be buying a house at this time.

  17. Hi David, Thanks for the note. It would be hazardous to create a default model without factoring in all variables–of which amortization is one. So before extended amortizations were launched in Canada, insurers modeled them with foreign data. There is ample such data available despite their short lifespan on high-ratio Canadian mortgages. And of course, (for now) 30-35 year amortizations are still available on new low-ratio mortgages. All this said, mortgages with longer amortizations do have a higher probability of default. It’s just not an overwhelming factor. From an underwriting standpoint, longer amortizations can easily be offset by other indicators of ability and willingness to pay. Cheers…

  18. I totally agree with you…leaving people hanging out to dry with high interest credit card debtand not able to refi to a low interest new mortgage…of course the Banks win if people do not refinance high interest credit card debt into low interest rate mortgages so who are they really protecting…just sayin..

  19. Hi Summa,
    Counterpoints are always welcome, thank you.
    When asserting that something is “risky,” it’s always helpful to quantify that risk. Otherwise “risky” is a completely unsubstantiated word.
    When it comes to extended amortizations on low-ratio mortgages, some of the risk-related considerations are:
    * The ratio of low-ratio borrowers choosing long-amortizations: 24% (Src: CAAMP)
    * The ratio of borrowers using long-ams to purchase a more expensive home: 24% (Src: Altus Group. Note: This figure reflects mostly high-ratio borrowers. It would likely be lower for low-ratio borrowers but we’ve yet to see that data.)
    * Increase in home prices due specifically to low-ratio extended amortizations: There are few, if any, studies on this but if I had to guess, it would be somewhere in the single digit percentages.
    * Average debt service ratio of a low-ratio buyer: Again, there’s little data specific to long-am borrowers but their gross debt service (GDS) ratio would likely not be considerably higher than the mid-20’s range for all insured borrowers (anything in the 20’s is very reasonable)
    In short, there seems to be little evidence that low-ratio amortizations are currently increasing risk in the housing market in any significant way. The bulk of the impact would have occurred after long-ams were first introduced. But since then, lending guidelines have been significantly tightened. And again, amortization length is only one of many factors affecting default risk and ability to service debt going forward.
    As for 60 or 100 year amortizations, low-ratio borrowers have–for a while now–had access to 100+ year effective amortizations through interest-only HELOCs. But, as usual, one factor alone (amortization) doesn’t determine how much the typical low-ratio borrowers pays for a home, and is qualified for.
    As for whether Canada is in a housing bubble, that’s something that others will continue to debate. Homes are undoubtedly overvalued in some—but not all—regions. Yet, most mainstream analysts and policymakers contest any and all claims of a bubble. Time will obviously tell.

  20. As long as the mortgage payment amount is similar to a rent amount, they should be good.
    Other expenses for an average home are approx 500$+ per month, so math is simple when planning the affordability in case of job loss.
    It’s another story if the person spends too much on a funnier lifestyle than making a rainy day fund :)

  21. One thing to keep in mind for these models is the interaction between pieces. A 30-yr vs 25-yr amortization is a pretty small piece of the default risk puzzle, but it interacts with interest rates: at 3% the extra 5 years of amortization can buy 12% more house at the same monthly payment whereas at 6% interest rates the extra amortization only gets a buyer 6% more house. If interest rates increase, the ceiling price for the same monthly payment drops 43% for someone on a 30-year loan vs 35% for someone on a 25-year one. If the model is built from data in a higher rate environment, it may understate the added risk of higher amortizations.
    Even having said all that, amortization is still a small factor. But if you’re trying to engineer a soft landing, then putting the focus on small factors at the margins may be just what you’re looking for.

  22. My wife and I each make in excess of 100k per year and could easily clear our mortgage in under 10 years but we signed a 40 year mortgage. Why? Flexibility and the ability to capitalize on record low interest rates to top up other investment vehicles and diversify. Why pay off my house and own one asset class when for 3%, I can top up tfsa, rsp, resp accounts? I am at zero risk for defaulting but according to the stats, I am one of the percentage points in a high ratio, long am mortgage.

Your email address will not be published. Required fields are marked *