OSFI Dismisses Renewal Impact of B-20

We’ve said it many times over. OSFI’s imposition of a stress test on borrowers switching lenders is potentially the most short-sighted government mortgage policy in Canadian history.

See: “Mortgage Renewals Now More Costly — For Those Least Able to Pay

The policy keeps borrowers—who have proven their ability to handle their mortgage—from switching lenders to reduce their interest bill. These are borrowers who already qualified previously and who would add zero new risk to the system simply by changing mortgage providers.

But OSFI has given them no relief. And last week, the banking regulator tried to justify its position by sending out this statement:

“Another concern was that B-20 would force a number of borrowers to remain with their current lender on renewal, potentially leading to an increase in their mortgage interest
rates. Although it appears more borrowers are staying with their original lender (renewals are up 30% [emphasis ours] while new mortgages are down 19% year-over-year in the April to July period), there has been no material change in the difference between the rates that renewal customers pay and those offered to new customers.”

That tells us little.

Keep Your Eye on the Ball

Of course there’s been little change for most renewal customers. “Most” renewal customers aren’t the concern because they can readily qualify at any lender. The problem is the meaningful minority of renewal customers who now cannot pass the government’s arbitrary new stress test. We estimate that number at roughly 1 in 10.

We asked OSFI how it could dismiss the impact on these renewers, given:

  • The stress test has made it impossible for a meaningful minority of borrowers to leave their lender for a better rate at another lender
  • Virtually every bank that publicly reports is reporting a material increase in customer retention rates (at some lenders a whopping 94 out of 100 borrowers are choosing, or having little other practical choice but, to remain with their lender at maturity).

OSFI responded:

“OSFI compared interest rates charged for renewals to interest rates charged for new originations, with further segmentation at the product level (e.g., 1-year fixed, 5-year fixed, variable rate, etc.). This was completed for a 24-month time period. This analysis is based on supervisory data and regulatory returns collected from federally regulated institutions.”

That’s great, but here’s the problem. The stress test is blocking borrowers who have a higher ratio of debt to income. (And note, those high debt ratios are not just reflective of borrowers with discretionary overspending. Some people do have unavoidable life challenges: medical crises, business failure, divorces and so on. They need lower rates more than anyone to remediate their finances.)

OSFI told us it could not measure the renewal rates paid by borrowers who had high total debt-service ratios (e.g., > 40%). It therefore did not compare such data from before and after the January 1, 2018, stress test introduction.

So how could the regulator know how “material” the problem really is? How could it state with confidence that its refusal to exempt mortgagors switching lenders is not harming a material number of Canadian families by elevating their interest burden? And what is “material” anyway?

Don’t Worry About Those Amortizations

OSFI also claims that “…The proportion of uninsured mortgages with amortization periods greater than 25 years has decreased from 51% to 47%, over the same April to July time period, suggesting that lenders are not extending amortization periods to allow borrowers to meet the stress-test requirements.”

While anecdotal, brokers and bankers we speak with are increasingly running into borrowers who need costlier extended amortizations to pass the stress test. So we asked OSFI, is it not possible that many of those who would have gotten extended amortizations can no longer qualify regardless (due to the stress test), and that this is a big reason why a lower proportion are getting amortizations over 25 years?

It replied: “The outcome you describe could be seen to improve the overall quality of mortgage originations by increasing the amortization of principal in the early years of a mortgage.”

Not sure that answered the question.

And let’s not forget, the cost of extended amortizations has risen since 2017 due to government regulation. Most lenders now charge 10 basis points more if you want an amortization over 25 years. Other things equal, higher prices result in lower demand.

Again, the people hit by OSFI’s B-20 are those with higher debt-service ratios. We then asked the regulator, “What percentage of those with high total debt-service ratios (e.g., > 40%) are getting extended amortizations today, versus before the stress test took effect in January 1?”

OSFI replied: “…OSFI collects supervisory data as part of our ongoing supervisory activities and the regulatory returns that we receive from regulated institutions. In the article, we provided high-level stats but we cannot go into the level of granularity you request.”

Hence, there is no clear visibility on how many borrowers are being required to take out longer amortizations because of the government’s stress test.

In its newsletter sent out to media and industry, OSFI goes on to state: “The federally regulated share of residential mortgages among regulated lenders in Canada for the 12-month period ending June 2018 remains stable (76.9% down to 76.7%).”

This says nothing about how many borrowers have had to move from prime lenders to non-prime lenders. So we asked OSFI, “What percentage of originations were done at prime federally regulated lenders (e.g., big banks) as of June 2018, versus non-prime federally regulated lenders (e.g., Home Trust, Equitable Bank and Equity Financial Trust)? How does this compare to before the Jan. 1 stress test?”

OSFI’s response: “OSFI is unable to provide the requested level of granularity, as noted above.”

Hailing Victory

In its press release (sorry, newsletter) OSFI said, “The proportion of uninsured mortgage originations that have loan amounts greater than 4.5 times borrower income declined from 20% in April-July 2017 to 14% in the same period in 2018.”

Clearly, our most systemically important lenders are seeing a better quality of borrow. No question about it.

But…OSFI’s data only reflects the proportion of high loan-to-income mortgages at federally regulated lenders. In other words, it doesn’t include the proportion now having to use higher cost credit unions, mortgage investment corporations and private lenders. Higher costs and worse terms increase the probability of default and/or insolvency. That spawns economic instability that indirectly adds risk for systemically important lenders.

The Bank of Canada tells us it “is looking at different data sources to try and get a better understanding of non-federally regulated lenders” but it can’t provide that information at the moment.

For these reasons, and due in part to the missing data, there are simply too many unanswered questions—so many, that drawing empirical conclusions about B-20’s impact on renewers is impossible. This is one reason why federal MPs like Tom Kmiec are asking the government to study the issue. Thus far, government has been unwilling to entertain that it might, just might, be wrong—and study the impact.

The federal opposition and trade groups, like Mortgage Professionals Canada, asserts that the federal government should make it a priority to get the needed data to confirm exactly how consumers are being negatively impacted by B-20’s renewal stress-test policy. And it should.

Policymakers claim the effect is not “material,” but claims that impact people’s lives should be supported with relevant data.

Sidebar: Making new borrowers prove they can afford a higher interest rate is undeniably vital. We—and most Canadians for that matter—are pro-stress test. The debate is only how to best administer a stress test with the least side effects.


  • Jim says:

    I’m not surprised that mortgage brokers are railing against OSFIs stress test: after all, it reduces new originations which hits their bottom line.

    I particularly love the line about “higher cost credit unions”. Many credit unions offer better products and rates than FRFIs – and lets be clear, the CU system is just as well regulated as the federal system!

    Shafting over extended borrowers (read: those that have high TDS and GDS) at renewal time is a good thing. I welcome repricing of their debt at market rates. Why should a borrower that has poor TDS/GDS be able to access low rates just because they’re already in the system? These borrowers pose a risk that needs to be addressed appropriately.

    • The Spy says:

      Jim, Broker opinions don’t change the facts. The fact are clear. Tens of thousands of renewers are now forced to pay more and families are therefore forced to incur more financial risk, for absolutely no good reason.

      On your second point, re-read the context. On average, mainstream CUs advertise lower rates than banks. But if you need to qualify at the contract rate, you don’t have a hope in hell of getting those rates. CUs charge those borrowers large rate premiums (50 to 75+ basis points), and rightly so. That is why CUs are “higher cost lenders” for such customers.

      Your personal agenda against renewing borrowers with higher debt ratios serves no purpose. For one, it fails to recognize the unavoidable adversity that leads many people to take on debt that they never would have taken on otherwise. Your lack of sympathy for these people doesn’t change the fact that forcing them to pay more, stay in debt far longer and defer retirement is an economic risk (to both them and the system). Furthermore, there is no evidence that imposing higher mortgage rates on already-indebted high-TDS borrowers is a deterrent factor that suddenly makes them “wake up” and shed their debt. What we do know is that adding to high-TDS borrowers’ interest burdens raises their default risk, which runs counter to any rational debt repricing objectives.

      As stated, renewing borrowers who have never missed a payment pose absolutely no additional risk merely by switching lenders. The regulated institution is going to renew them regardless and therefore the “risk” stays in the system regardless. There is simply no good result that can come from making life more expensive for individual Canadian borrowers unless doing so reduces system-wide risk and materially benefits the economy at large. And in this specific case, it most certainly does not.

  • Jim says:

    There is a good reason that those tens of thousands of renewers should be paying more: they are heavily indebted and have high TDS/GDS! If a corporate borrower gets downgraded and subsequently has to reissue debt to pay for upcoming maturities: the price becomes higher. Let the market price the risk accordingly!

    You’re spot on TheSpy: I have little sympathy for highly indebted borrowers. Allowing those borrowers to stay in the system paying the same price as the rest of us prudent borrowers is offensive to me (and should be for everyone involved).

    As someone involved with CU loan underwriting, I believe, generally, that Canadian underwriting is sound. As we all know, 200bps on a $500k loan is about ~$500 a month. I don’t believe that many borrowers are $500 a month away from default! As our esteemed head of CMHC likes to say: “the biggest risk to the financial system is unemployment”. I would agree, repricing these loans (ie: making switch renewers subject to the stress test) won’t, on its own, cause a systemwide meltdown. Now, if we went to nationwide 10% unemployment, we’d have a different discussion. That said, in that case I’m pretty sure the overnight rate would get close to zero PDQ!

    As I’ve illustrated, I believe loan pricing (for residential mortgages) is an issue. Generally lenders deal with absolutes: a borrower with TDS at 42% gets the same price on their debt as a borrower with a TDS of 1%. I don’t think that’s fair, nor do I think it’s prudent. Personally, I would go so far impose the stress test on all renewals! Corporates are subject to that when their MTNs and CP issues come due, why shouldn’t individuals? The market will adapt: people who are worried about their debt repricing in 5 years will take on 20 year terms. Again – this is something a corporate treasurer has to worry about, right? Why should the treasurer of Joe Public’s household avoid that concern?

    Good debate! I’m glad you published my comments!

  • Sam McGee says:

    +1 Spy

    The guy who wants to “shaft” borrowers sounds like a greedy lender, an uninformed housing bear or a regulator trying to save face.

  • Jim says:

    Come on Sam McGee – is there any need for that? Can you explain why someone with TDS of 42% gets the same rate as someone with TDS of 1%? Why is that OK to you? The risk profile of those two borrowers is very different, isn’t it? Please, educate this “greedy lender, uninformed housing bear and/or regulator”.

  • Guy says:

    LOL. Jim implies mortgage brokers are biased on this issue and then goes on to say he’s a lender and wants to charge people higher rates. I’m sure he also loves how OSFI has guaranteed lenders sky high retention rates. No impact on the “bottom line” there, hey Jim? Pot-kettle.

  • Jim says:

    I don’t need B-20 revisions (which haven’t applied to us until recently) to accurately price our loans. I didn’t need B-20 changes to allow our CU to “charge higher rates”. If a borrower is highly indebted and over levered, they pay a rate premium (now and before!).

    B-20 hasn’t guaranteed our CU anything: we haven’t been blindly renewing borrowers without updating our underwriting. If their risk profile changed then their rate changed! How does B-20 change this?

    I’ve asked this question 3 times now but nobody seems to have an answer: why should loan pricing for a 42% TDS borrower be the same as the pricing for a 1% TDS borrower? Isn’t the risk profile different?

    • The Spy says:

      Hey Jim,

      First off, know that your posts are appreciated. While there may be disagreement, an honest debate is the fastest path to the truth.

      Where there is no debate is that borrowers with materially higher debt loads relative to income should be risk-priced at origination. Could not agree with you more in this regard.

      Renewal is different. The borrower already has the mortgage. The lender’s money is already at risk.

      At renewal, the lender’s #1 concern making sure the borrower can pay them back.

      Layering on rate premiums does not raise the probability of repayment. It increases default risk, particularly with higher-TDS borrowers in a falling housing market.

      The argument that rate premiums afford lenders capital buffers ignores the facts that doing so:

      * Reduces their probability of repayment
      * Does not endear borrowers to the lender, which presents future business ramifications
      * Prevents competitive lenders from getting renewal business
      * Causes financial damage to families, despite that 99% +/- of such borrowers will repay without issue.

      OSFI states: “FRFIs are not expected to re-apply the qualification rate assessment to existing borrowers that are renewing mortgages.” There are practical reasons for that. For one, OSFI recognizes that it creates risk to the lender (let alone the borrower) to strand borrowers at renewal.

      Back when it formulated B-20 OSFI stated: “Current practice regarding residential mortgage renewals has served FRFIs well. OSFI agrees, for example, that having a good payment record is one of the best indicators of credit worthiness.”

      While underwriting at origination, the institution hasn’t committed capital yet and does not have its own track record of the borrower. So there’s more cause to ensure a borrower can afford future rate hikes. At renewal, given 20%+ equity.

      Of course, OSFI does require lenders to “refresh the borrowers’ credit metrics periodically (not necessarily at renewal) so that FRFIs can effectively evaluate their credit risk.” And that’s fine. Testing 5-year fixed borrowers at the contract rate at renewal accomplishes that (qualifying at contract is not appropriate for some terms, like 1-years or adjustable rate mortgages). Remember that for decades, prime lenders tested ALL 5-year fixed borrowers at the contract rate and the system was rock-solid with ultra-low arrears. In this context, we’re only talking about renewers with proven repayment track records. And for greater safety, transfers qualified at the contract-rate can be limited to 5-year fixed terms–similar to the government’s former policy on insured mortgages.

      In sum, the renewal borrower’s risk profile is clearly different than a brand new purchaser or refinancer. And that’s why the stress test exception is warranted.

  • Jim, I like the idea of pricing according to risk, but the devil is in the details. Someone with a 40% TDS, $100,000 net worth, and a 750 beacon score could be lower risk than someone with a 35% TDS, $5,000 net worth, and a 680 beacon score.

  • Leslie says:

    All things considered, this new rules holds reliable borrowers hostage to their existing lender, leaving lenders free to issue existing borrowers unreasonably higher renewal rates. Simply because the lender knows the borrower is in fact their hostage. How is this legal? How is this not a monopoly over the existing borrowers. Under the new rules why wouldn’t the existing lender require an unreasonably higher rate to renewals to increase their bottom line? Competition keeps lenders as honest as possible. Now what incentive is there to reward reliable and excellent paying borrowers with competitive rates?

  • Leslie, about a month ago I was offered prime -.86 to early renew into a 5yr variable at Scotia. That’s lower than the prime -.75 that RateSpy had listed for Scotia, and certainly couldn’t be called “unreasonably higher renewal rates”.

    • The Spy says:

      Two comments on your post Ralph.

      First, just so everyone is clear, the rate you cited for Scotia is an estimate of the average Big 6 bank’s discretionary 5-year variable rate. Individual banks will vary, as will the experiences of individual clients. The discretionary rate estimate is meant as a starting point for negotiations with the bank (i.e., the worst rate you should expect on an owner-occupied standard 25-year mortgage–if you’re reasonably qualified). In many cases, customers will get a better rate. In some cases, they’ll get an inferior rate.

      Second, the renewal impact of B-20 is real but only for a minority of borrowers (roughly 1 in 10 we estimate). Hence, one example doesn’t establish much. If we’re talking anecdotes, I had one B-20-compliant customer renew at an exceptional prime – 1.00% this summer. By contrast, we currently have a non-B-20-compliant bank client who just got quoted a lousy 3.99%. We’re getting him done at a credit union at 3.69% (no OSFI stress test).

Leave a Reply

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