Introduction
Since March 2022, interest rates have risen considerably and rapidly following a period of historical low rates during the first two years of the COVID‑19 pandemic. As a result, many mortgage holders are currently facing significantly higher payments, and others will do so at renewal. The exact size of this increase in payments depends on the features of each mortgage and how interest rates continue to evolve.
To assess how interest rates could further impact the cost of servicing mortgages, we use loan-level data to simulate future mortgage payments under the assumption that interest rates evolve according to financial market expectations.1 Thus, this simulation is hypothetical and does not represent a forecast.
We find the following:
- By the end of November 2023, about 45% of the mortgages taken out before the Bank of Canada started raising its policy interest rate in March 2022 had seen an increase in payments. By the end of 2026, virtually all remaining mortgage holders in this group will go through a renewal cycle and, depending on the path for interest rates, may face significantly higher payments.
- Borrowers who either took out a mortgage in 2021—when interest rates were at historical lows—or opted for a variable mortgage rate will generally have experienced the largest increases in payments by the end of 2026. Among variable-rate mortgage holders, those with fixed payments who have not taken action to avoid large future increases will be impacted at renewal. For this group, median payments are expected to increase by 54% during the period between the end of February 2022, just before interest rates began to increase, and the end of 2027. In contrast, those with variable payments have already been impacted, with median payments up 70% in November 2023 compared with their level at the end of February 2022. However, based on market rate expectations, payments are expected to decline for this group starting in mid-2024.
- The impact of higher interest rates on borrowers’ ability to pay their mortgage will largely depend on their future income. Without any income growth, the median borrower may need to dedicate up to 4% more of their pre-tax income to mortgage payments by the end of 2027. However, for some borrowers, income growth could mitigate the impact of higher interest rates on debt serviceability.
It is important to note that our simulation does not account for potential changes in the behaviour of borrowers, such as making accelerated payments or switching to another mortgage product. Such changes would help lessen (although not avoid) the increase in payments. Thus, our simulation results represent an upper-bound estimate.
The dataset
Our simulation uses anonymized, regulatory, loan-level data collected by the Office of the Superintendent of Financial Institutions (OSFI), Canada’s banking regulator. Microdata compiled by OSFI have the most comprehensive information available to analyze the impact of interest rates on mortgage payments in Canada, allowing for the most accurate and granular simulation possible. In the dataset used for the simulation, we observe each mortgage at origination (for either a new purchase or a mortgage refinance) and at renewal. The dataset includes about 16 million mortgage observations since 2014. A few important notes on this dataset:
- It includes mortgage activity at federally regulated lenders, including the Big Six Canadian banks and smaller banks. Mortgages at other types of lenders, such as credit unions and mortgage financing companies (MFCs), are generally not included in this dataset because these lenders are not regulated by OSFI.2 Our dataset therefore covers about 80% of the total mortgage market.
- It captures various loan and borrower characteristics, including the:3
- initial (contractual) interest rate
- income used in the application to qualify for the mortgage
- size of the loan
- contractual amortization period
- loan term
- type of interest rate—fixed or variable4
- It is a “flow” dataset, meaning that we observe each mortgage only at the initial purchase or when the mortgage is refinanced or renewed. In between these occurrences, we simulate the paydown schedule of each mortgage.5
The simulation exercise
For each mortgage in our dataset, we calculate the mortgage payment over history and then simulate the future payments based on an assumed path for interest rates. We construct the interest rate path using historical mortgage rates combined with expectations derived from financial markets for both the policy interest rate and government bond rates. As shown in Chart 1, in mid-December 2023, financial markets were expecting the policy interest rate to peak at the end of 2023 and then remain higher than it was on average over the years before the pandemic. We use the path as a benchmark for the rate increase each individual mortgage will face.6
Chart 1: Participants in financial markets expect interest rates to remain higher than they were before the COVID-19 pandemic
We then account for different paydown schedules depending on the type of mortgage.
- Fixed-rate mortgages: The mortgage payments are fixed for the duration of the loan term. The principal and interest portions evolve according to a regular amortization schedule. At renewal, the interest rate is updated, and the payment is recalculated.
- Variable-rate mortgages with variable payments: The interest rate is updated each month following changes in the prime lending rate, which is assumed to move in step with the policy rate. The outstanding balance and mortgage payment are then recalculated each month using the updated interest rate and remaining amortization period.
- Variable-rate mortgages with fixed payments: The interest rate and the corresponding interest portion of the mortgage payment are updated each month in line with any changes in the prime lending rate. However, the mortgage payment itself remains the same unless the mortgage reaches its trigger point (see Box 1). At renewal, the monthly payment is recalculated to bring the mortgage back to its original amortization schedule.7
We assume that at the end of the loan term each mortgage holder renews into the same type of product (i.e., a variable-rate borrower once again selects a variable rate).8 We also assume that borrowers do not make other changes, such as refinancing a mortgage or making prepayments. Overall, these assumptions likely lead us to overestimate the size of payments at renewal because some borrowers will take action to avoid a future large increase in payments.
Box 1: The case of variable-rate mortgages with fixed payments
Box 1: The case of variable-rate mortgages with fixed payments
In Canada, variable-rate mortgages can have payments that are either variable or fixed, depending on the lender. Fixed payments are the most common, comprising about 75% of variable-rate mortgages.
When the prime lending rate changes, variable-rate mortgages with variable payments are immediately subject to a change in payment size. But for variable-rate mortgages with fixed payments (VFMs), only the interest portion of the payment changes. For these mortgages, when interest rates go up, a larger share of the payment goes to interest, but the overall payment amount remains the same.
As discussed in Murchison and teNyenhuis (2022), the trigger rate is the interest rate at which the mortgage payment equals the interest payment. As a result of the large increases in interest rates that have taken place since early 2022, some VFMs have reached a point where the entire mortgage payment goes to the interest portion (and none to the principal portion). We estimate that by the end of November 2023, up to 80% of VFMs at federally regulated lenders had reached their trigger rate. This is an upper bound based on the trigger rates set out in the mortgage contracts. In reality, many VFMs have likely avoided reaching their trigger rate by making prepayments or converting to fixed-rate mortgages.
The trigger point, meanwhile, is determined by the lender and described in the mortgage contract. It is the point at which borrowers are required to increase their payment to cover interest. Some lenders require borrowers to increase their mortgage payment as soon as they reach the trigger rate. Other lenders, however, will allow the interest shortfall to be added to the balance of the loan up to a certain threshold. In this case, the balance owed on their mortgage grows over time as the interest shortfall accumulates—a situation known as negative amortization.
We estimate, based on lender-specific policies, that by November 2023, at most one-quarter of VFMs had reached their trigger point and been subject to a mandatory change in payment. Even if the policy interest rate that prevailed in October 2023 were maintained for several years, most VFMs would not reach their trigger point.
Many VFMs are currently in a situation of negative amortization; as a result, their loan balances are growing. Unlike a variable payment arrangement in which the payment always adjusts with the interest rate, these mortgages will face a larger increase in payments at renewal to get back to their original amortization schedule. Ultimately, they will end up costing more in interest over the life of the loan.9
The results
Many borrowers have yet to see an increase in their mortgage payments
We estimate that, among the stock of mortgages outstanding in February 2022, about 45% had already been subject to a payment increase by the end of November 2023, and 80% will face a payment increase by the end of 2025 (Chart 2). These figures do not include mortgages that were originated after February 2022.10
Chart 2: Most mortgages have not yet seen an increase in payments
Based on the simulation results, we expect the median increase in mortgage payments to be relatively large.
- Under our scenario where interest rates evolve according to financial market expectations, the median monthly mortgage payment for all outstanding mortgages will increase from $1,200 in February 2022 to $1,600 by the end of 2027—an increase of 34% (Chart 3, panel a).
- Variable-rate mortgage holders with variable payments are affected immediately by interest rate increases. But as financial markets expect interest rates to begin decreasing in 2024, payments also moderate slightly by the end of 2027.
- In contrast, variable-rate mortgage holders with fixed payments initially keep the same payment (unless they reach their trigger point, as discussed in Box 1). But they will face a larger increase in payments at renewal to return to their original amortization schedule. As many variable-rate mortgages with fixed payments are scheduled to renew in 2026 or 2027, the median payment for this mortgage type increases sharply in those years, reaching $2,190 by the end of 2027—an increase of 54% from the February 2022 level.
- If we assume instead that mortgage rates remain at their October 2023 level rather than evolve according to market expectations, payments will rise even further. In this case, the median mortgage payment for all outstanding mortgages increases 44% by the end of 2027 relative to February 2022 (Chart 3, panel b).
Chart 3: Median payments will continue to increase as more mortgages come up for renewal
Chart 3: Median payments will continue to increase as more mortgages come up for renewal
Median monthly mortgage payment for outstanding mortgages
Sources: Regulatory filings of Canadian banks and Bank of Canada calculations, estimates and projections
Last data plotted: December 2027
The impact on debt serviceability will depend on income growth
To provide a more complete picture of the impact of the rising interest rates on mortgage holders, we also need to consider what happens to income. Because household income tends to rise over time, some borrowers may see only a modest increase in their mortgage payments relative to income. Others may experience smaller income gains or a reduction in their income that could challenge their ability to meet their mortgage payments.
To examine this issue, we express mortgage payments as a share of pre-tax income for each mortgage in the sample. This indicator is called the mortgage debt service ratio (MDSR).
In our dataset, the only information available on income is the income used for mortgage qualification, which is collected by lenders when the mortgage is originated. We must therefore make assumptions about the future path of income. We consider two cases:
- No income growth: First, we assume no income growth between when the mortgage was originated and its renewal. In this case, the median MDSR for all outstanding mortgages would increase by 4 percentage points, from 16% in February 2022 to 20% at the end of 2027.
- Average income growth: Alternatively, we assume that for each mortgage holder in the dataset, income grows by 2.4% per year. This growth rate represents the average change in hourly wages in Statistics Canada’s Labour Force Survey over the period from January 2014 to September 2023. When we factor in this simple assumption of income growth, the median MDSR across all mortgage types increases by 1.5 percentage points from February 2022 to the end of 2027.11
Income growth dampens the impact of rising interest rates on debt serviceability, but the increase in the MDSR would still be still historically large given the extent of interest rate increases. Of course, not all mortgages are affected in the same way. The impact depends on the type of mortgage and timing of origination:
- Overall, mortgages that were originated in 2021—when interest rates were very low—and variable-rate mortgages will see the largest increases in debt-servicing costs at renewal compared with at origination.
- Variable-rate mortgages with fixed payments face an especially large increase because many have been negatively amortizing and thus have seen their loan balances grow. For instance, when we factor in the historical average of income growth, the median borrower who took out a mortgage in 2021 with a variable interest rate and fixed payments will see their MDSR increase to 22% at renewal, an increase of 5 percentage points compared with at origination (Chart 4).
- In the absence of income growth, the effect would be much larger. Variable-rate, fixed-payment mortgages originated in 2021 could face an increase of 8 percentage points at renewal compared with at origination.
Chart 4: Income growth would reduce the increase in the mortgage debt service ratio but would not fully offset payment increases
In reality, the impact of interest rate increases on debt serviceability will differ greatly for each borrower. Younger homeowners tend to have higher income growth than the median household. However, some borrowers may experience a reduction in income or a personal situation that challenges their ability to continue meeting their debt obligations. In addition, inflation has pushed up the prices of goods and services, reducing the purchasing power of households. To counteract this, some borrowers may be able to smooth the impact of payment increases by drawing on savings or home equity.
Overall, most borrowers will need to make adjustments, sometimes significant, to ensure they are able to continue to pay their mortgage. The impact on financial stability, however, is somewhat mitigated by macroprudential policy. For instance, when they first take out a mortgage loan, most borrowers have to undergo a mortgage rate stress test in which they demonstrate that they could afford their mortgage at a hypothetical higher interest rate.12 Under the current guidelines, new borrowers are stress-tested at an interest rate at least 2 percentage points higher than their contractual mortgage rate. This should provide a buffer for borrowers to absorb interest rate increases and other shocks. That said, the buffer will not be sufficient to fully offset the impact of large interest rate increases for some borrowers, even without other shocks such as a decline in income. We estimate that the median renewal rate for mortgages originated in 2020 will surpass the rate at which they were stress-tested by about 0.3 percentage points. For other borrowers, the median renewal rate generally does not surpass the rate at which they were stress-tested.
Conclusion
In this note, we investigate the extent to which the large increase in interest rates since March 2022 impacts mortgage payments.
We find that the majority of mortgages outstanding in early 2022 have not yet renewed at higher interest rates, but they will do so over the coming years. Overall, the ability of borrowers to service their mortgage will depend largely on the size of their mortgage payment relative to their disposable income.
One key implication of this work is that as long as they continue to experience income growth, most mortgage borrowers will not face severe financial stress from the increase in mortgage payments over the coming years. However, borrowers who stretched to enter the market or who were anticipating rate decreases by the time of renewal may find the adjustment more difficult. Moreover, a severe recession causing higher unemployment could challenge the ability of many to continue making their mortgage payments. This could lead to credit losses for lenders if mortgages exceed property values. In turn, it could also tighten lending conditions, making it more difficult and expensive for Canadian households to access credit.
The Bank will continue to closely follow the evolution of measures of debt serviceability as part of its assessment of risks to financial stability.
Endnotes
- 1. This note complements the analysis published in Box 1 of the Financial System Review—2023 (Bank of Canada 2023).[←]
- 2. The dataset also includes mortgages sold to large banks by MFCs. The Big Six Canadian banks are the Bank of Montreal, the Bank of Nova Scotia, the Canadian Imperial Bank of Commerce, the National Bank of Canada, the Royal Bank of Canada and the Toronto-Dominion Bank.[←]
- 3. This dataset does not include the size of the mortgage payment, but this can be estimated using other variables in the dataset and assuming monthly payments with semi-annual compounding of interest.[←]
- 4. The data do not specify whether a variable-rate mortgage has fixed or variable payments. However, as lenders generally offer only either variable or fixed payments, we assign the payment type based on the lender. Variable-rate mortgages issued by MFCs, which are included in our dataset if they are sold to large banks, generally have variable payments. We assign all MFC variable mortgage rates as having variable payments in our simulation.[←]
- 5. OSFI is now collecting an upgraded version of this dataset that provides a snapshot of every mortgage outstanding each month (“stock” concept), while the original dataset includes only monthly information on originations or renewals (“flow” of new mortgage lending). The upgraded dataset thus facilitates monitoring of actual loan amounts over time by capturing prepayments. These data will be incorporated into our simulation in the near future.[←]
- 6. The interest rate of each individual mortgage will follow observed trends in average mortgage rates to the last observation, then it will follow trends in financial market expectations for the overnight rate (for variable-rate mortgages) or Government of Canada bonds (for fixed-rate mortgages). The risk premium associated with each mortgage is reflected in the original mortgage rate and is held constant throughout the simulation.[←]
- 7. For instance, if a mortgage has a 30-year contractual amortization at origination, at the end of a 5-year loan term, the payment will be recalculated using the remaining 25-year amortization. Our simulation thus assumes that borrowers do not extend their amortization at renewal. A separate concept, the effective amortization period, specifies the theoretical amount of time it would take to pay down a mortgage at the current rate of payments. For variable-rate mortgages with fixed payments, the effective amortization will increase automatically when variable mortgage rates rise. At renewal, however, the effective amortization will also be reset along with the recalculation of the fixed payment. For more information, see Office of the Superintendent of Financial Institutions (2023).[←]
- 8. If we observe in the dataset a mortgage renewal that we can link to a previous observation, we use the new observation to update the previous mortgage. Thus, if the borrower renews a few months early or makes any prepayments, we capture that when we add in the new observation. Mortgages in the past (before our data ends) are removed from the constructed stock at the end of their loan term because we should observe a renewal or refinance of that mortgage at or before the end of the loan term, even if it cannot be explicitly linked to a previous observation.[←]
- 9. For example, a VFM originated in January 2022 that permits negative amortization would cost 6% more over a 30-year amortization period than a mortgage with variable payments, assuming interest rates evolve in line with financial market expectations until the end of 2027 and are then constant after 2027. For a $600,000 mortgage, this means an additional $70,000 in interest payments over the life of the mortgage. [←]
- 10. After interest rates began increasing in March 2022, variable-rate mortgages decreased in popularity, and new borrowers tended to opt instead for short-term fixed mortgage rates (primarily three years). Thus, most new mortgages will face an increase in payments at renewal.[←]
- 11. The MDSR is estimated as a share of qualifying income, which is a pre-tax (gross) measure. If an after-tax (net) measure of income were used instead, the increase in MDSR would be larger.[←]
- 12. New mortgages originated after June 2021 must satisfy debt service ratio requirements when mortgage payments are calculated using the greater of 5.25% or the contract rate plus 2 percentage points. Similar guidance applies for most mortgages originated before June 2021. For more information on mortgage rate stress tests, see, for example, Bank of Canada (2018) and Office of the Superintendent of Financial Institutions (2021).[←]
References
Bank of Canada. 2018. “Box 1: Mortgage interest rate stress tests.” Financial System Review (June): 5.
Bank of Canada. 2023. “Box 1: The impact of higher interest rates on mortgage payments.” Financial System Review—2023 (May).
Murchison, S. and M. teNyenhuis. 2022. “Variable-Rate Mortgages with Fixed Payments: Examining Trigger Rates.” Bank of Canada Staff Analytical Note No. 2022-19.
Office of the Superintendent of Financial Institutions. 2021. “Backgrounder – Minimum Qualifying Rate.” (December 17).
Office of the Superintendent of Financial Institutions. 2023. “OSFI’s View: Variable Rate Mortgages with Fixed Payments and Extended Amortizations.” (September 13).
Disclaimer
Bank of Canada staff analytical notes are short articles that focus on topical issues relevant to the current economic and financial context, produced independently from the Bank’s Governing Council. This work may support or challenge prevailing policy orthodoxy. Therefore, the views expressed in this note are solely those of the authors and may differ from official Bank of Canada views. No responsibility for them should be attributed to the Bank.
DOI: https://doi.org/10.34989/san-2023-19