Economic insight provided by Alberta Central Chief Economist Charles St-Arnaud
- House prices have increased sharply over the past year and comes on the back of a multi-year appreciation. This raises concerns regarding whether the housing market could be in a bubble.
- We look at valuation and affordability in Canada’s main cities: Vancouver, Calgary, Edmonton, Winnipeg, Toronto, Ottawa and Montreal.
- One of the main findings is that the current record low interest rate environment is a powerful force keeping affordability high, despite continued increases in house prices.
- Moreover, our analysis shows that Toronto is the most overvalued city in Canada, followed by Ottawa, Montreal and Vancouver. At the other end of the spectrum, there are no signs of overvaluation in Edmonton, Winnipeg and Calgary.
- With record low interest rates being the main source of affordability in most cities, an important question for the housing market is how valuation will be impacted by higher interest rates?
- Our simulations show that many cities in Canada will struggle with housing affordability as interest rates increase, 150bp increase in mortgage rates could be enough to generate significant headwinds on some housing markets and house prices.
- The high sensitivity to interest rates also supports the case for a gradual reduction in monetary stimulus and suggests the peak in the interest rate cycle will likely be lower by historical standards.
- We currently expect the Bank of Canada to start increasing its policy rate in 2022Q3.
House prices across Canada have increased sharply over the past year despite the economic downturn and the high unemployment level resulting from the COVID-19 pandemic. This situation comes on the back of a multi-year appreciation, where prices have almost doubled over the past decade and are four times higher than 20 years ago. The continued increase in house prices leads to an intensification in concerns regarding affordability and whether the Canadian real estate market is in a bubble.
In this report, we look at various house price indicators to better assess the valuation of the housing market. The aim is not to determine whether the market is a bubble. Instead, we are looking at affordability, whether it is becoming stretched, and if future conditions could spell trouble for the housing market.
It is of little relevance to study the housing market in Canada by taking a national perspective, as it is too costly or a very low occurrence for someone to buy in a region other than where they work and commute regularly. Instead, the national housing market should be viewed as a collection of local markets influenced by national policies (monetary policy, financial policy, CMHC rules, etc.), but determined by regional evolutions (population, income, supply, etc.). As a result, to get a good sense of the valuation in the housing market, one needs to look local metrics.
For this reason and because of the data availability, our analysis focus on the main metropolitan areas of Canada: Vancouver, Calgary, Edmonton, Winnipeg, Toronto, Ottawa and Montreal. Altogether, those seven cities cover about 50% of the Canadian population.
The following valuation metrics are used: house price relative to income, house price relative to rent, mortgage payments relative to income and mortgage payments relative to rent (see appendix for detailed results).
Whether an asset is overvalued or not is often subjective as there are no defined levels and this exercise relies heavily on the difference between current prices and “normal.” For this reason, we also look at the deviation of the various metrics relative to their historical mean using the Z-score (see Appendix for background on Z-score).
This point is important because, as mentioned previously, housing markets are local and depend on local conditions. For example, conditions in Vancouver have meant that prices have generally been higher than elsewhere in the country for the past four decades. Yet, someone living in the city knows and understands that this is the reality of living there and will plan accordingly. Someone in Vancouver is unlikely to consider a house in Winnipeg or Edmonton, even if the real estate is much cheaper, because commuting is practically impossible due to the commuting cost.
Key valuation findings:
- There is a clear difference between measures that include interest rates and those that do not. The average Z-score for house prices relative to income suggests overvaluation in many cities. However, when considering mortgage payment relative to income, the Z-score drops to close to 0 on average, meaning that affordability is not an issue, at current interest rate levels.
- Across the various valuation measures, Toronto comes in as the least affordable city in our sample, followed by Ottawa and Vancouver.
- At the other end of the spectrum, Edmonton was the most affordable city, followed by Calgary and Winnipeg.
- The case of Vancouver and Ottawa are interesting. All of Vancouver’s valuation metrics in absolute terms are the highest in Canada. However, relative to historical mean, these levels are closer to their averages, suggesting only stretched valuation. While absolute measures are closer to the national average in Ottawa, they are considerably elevated compared to history, often suggesting deep overvaluation.
- Toronto has both elevated measures in absolute terms and relative to history, making it the least affordable city in our analysis. Moreover, when when only considering the metrics that includes interest rates, valuation in the city is stretched.
These findings show that record low interest rates are keeping houses affordable in many cities. This situation is particularly acute in Toronto, Vancouver, Ottawa and, to a lesser extent, Montreal. It raises concerns regarding what will happen to the housing market when interest rates inevitably increase.
It also points to a situation where the biggest hindrance to affordability at the moment is the higher level of the down payment relative to income. As such, an average family in Toronto would need to save two-thirds of it annual income to accumulated the minimum down payment of 10%. This proportion is about three-quarters of annual income for Vancouver.
As we showed, interest rates are the main reason why affordability in many cities in Canada remains close to their historical averages. However, we are currently at the bottom of the rate cycle, meaning interest rates are likely to increase in the coming years. Therefore, an important question for the housing market is how much can we expect affordability to be impacted by higher interest rates?
Using the metrics used previously (and outlined in Appendix), we conduct a series of simulations to estimate the impact of rising interest rates and changes in house prices on the valuation measures. Since our aim is mainly to evaluate the effect of interest rate hikes, we will focus on the metrics of mortgage payment relative to income and mortgage payment relative to rent.
The simulations are conducted by varying both the interest and house prices, allowing us to show the combined impact of both variables. It also provides an estimate of how much house prices would need to decline to compensate for the reduction in affordability coming from higher interest rates. All the simulations leave the level of income and rent unchanged from their latest values.
The higher interest rates simulation could also be used to assess the impact of the mortgage stress test on affordability. However, the higher interest rate used in the stress test is only to determine whether a borrower could face higher interest rates and does not impact the actual mortgage payment of the borrower. As such, it has a smaller impact than an actual increase in interest rates.
Key simulation results:
- Focusing on mortgage payment relative to income, on average in Canada an increase of 400bp in mortgage rate would be required for the average Z-score in Canada to reach 2, or an overvalued level, assuming house prices remain constant. If prices were to increase by 10%, it would require a 300bp increase in rates. If prices rise by 20%, an increase of 200bp would be required.
- However, the average masks some stark divergences between cities. For example, there are very few concerns for Calgary, Edmonton and Winnipeg. In all three cases, even a 500bp increase in interest rates would leave the measure of mortgage payment relative to income well below the overvalued threshold. However, for overvalued cities like Toronto, a 150bp increase in rates would be sufficient to push the Z-score above 2 at current prices, 100bp if prices increase 5% and 50bp if prices increase by 10% from current levels. A 15% rise in house prices would push the market into the overvalued territory at current interest rates.
- If we only consider the average for Toronto, Vancouver, Montreal and Ottawa, the most overvalued cities representing almost 40% of the Canadian population, at current prices, a 250bp increase would push those markets into overvalued territory. If prices rose by 5%, those markets would be overvalued after a 200bp hike and a 150bp hike if prices increase by 10%.
- If we consider the mortgage-to-rent measure, all markets would present signs of overvaluation sooner. This suggests that even with the current low interest rate, mortgage payment are expensive compared to rent.
- Based on mortgage-to-rent, a 150bp hike in interest rates would push the average of all the cities into overvalued territory. If prices increase by 5%, the mark would be reached with an increase of 100bp, and 50bp if prices go up by 10%. In addition, at the current level of interest rates, a 15% increase in house prices would push the average city into overvalued territory.
- Removing Calgary, Edmonton and Winnipeg, the least overvalued cities from the sample, leads to significant changes. As such, the average of the remaining cities suggests they are already in the overvalued territory at the current level of prices and interest rates.
- However, as mentioned previously, it is important to note that since a part of the mortgage payment stays with the homeowner as equity, the comparison may not be the most accurate in the current record low interest rate environment.
- If mortgage interest payment relative to rent is considered, a 150bp increase in interest rates would push the average measure for all cities into the overvalued territory and a 100bp rise would do the same, if only considering Toronto, Montreal, Vancouver and Ottawa.
The main point from those simulations is that many cities in Canada will struggle with housing affordability as interest rates rise and those increases don’t need to be big. A 150bp increase in mortgage rate could be enough to generate significant headwinds on some housing markets and house prices.
The main conclusion from the simulation is that, while rates will eventually increase, the speed of the normalization and the level that rates will reach is likely to be low by historical standards. This is because policymakers will be concerned about a sharp decline in the housing market given the ripple effect it could have on household borrowing and financial institutions.
Moreover, higher interest rates will also impact households’ budgets by increasing their mortgage payments for those with flexible rate mortgages or those renewing their borrowing. While this situation may not lead to an increase in insolvencies, unless the rate increase is sizable, some households will likely force some households to restrain on spending.
The high sensitivity to interest rates also suggests that there are some benefits for the central to being pre-emptive in its removal of monetary policy stimulus, in a effort to avoid having to increase rates at a more rapid pace at a latter date. This would allow the rate sensitive sectors to more gradually adapt to the new environment.
To provide some context, the Bank of Canada’s policy rate currently sits 150bp below its pre-pandemic level and mortgage rates are about 135bp lower than their peak in 2019. However, interest rates were still not normalized at that time. The Bank of Canada believes its neutral rate sits between 1.75% and 2.75% or 150bp to 250bp higher than the current level. As a comparison, before the global financial crisis of 2008, the BoC’s policy rate was 425bp higher and mortgage rates were about 355bp higher than currently.
The valuation metrics and simulations also suggest another potential economic development in the coming years: rents have the potential to increase sharply.
- As interest rates go up and households become increasingly priced out from the housing market, demand for rentals is likely to increase, putting upside pressures on rents.
- Rental prices relative to income ratios in many cities are below their long-term average. This is partly the result of rental prices under performing over the past year, as many renters decided to leave cities or purchase homes during the pandemic. However, some mean-reversion could be expected, especially as the cost of owning will be pushed higher by increasing interest rates.
The amount of pressure on rental cost will also depend heavily on demographics, especially immigration, and on the supply of new rental units.
The valuation metrics and the simulations show that the current low interest rates are a significant driver for the housing market by keeping affordability high despite elevated and increasing house prices. However, interest rates are likely to increase soon, with the Bank of Canada expected to hike its policy rate in the second half of 2022.
Our simulations show that a rise in mortgage rates doesn’t need to be significant before some headwinds on the housing market begin to be felt. For example, in many cities, a 100bp increase could push the valuation metrics into overvalued territory. However, this doesn’t mean that we will see prices collapse, rather that this will likely lead to a housing market stagnation.
This situation will have an impact on monetary policy. It will likely mean that any rate increases will be gradual to allow the housing market to adjust and that the terminal rate will be lower than in previous rate cycles because of the negative feedback loop between rates, mortgage payments and discretionary spending.
The Z-score measures how far away from the historical average the observation is in terms of standard deviation. As such, a Z-score of 0 indicates that the observation is on the historical average. A negative score means that it is below, while a positive score indicates it is above.
Moreover, the property of the Z-score shows how likely such an observation is, based on history. The chart below shows that about 68% of all the historical observations fall between a Z-score of -1 and +1 and 95% have a Z-score between -2 and +2. This is useful, as it means that a Z-score above 2 only happens 2.5% of time. In our current analysis, this means that a metric with a Z-score of higher than 2 only happened 12 times between 1980 and 2021, based on 495 observations.
Using those properties of Z-score, we defined a metric with a Z-score between 1 and 2 as having a stretched valuation. Once the z-score is above 2, we will define the measure as being overvalued. It is important to keep in mind that, using the Z-score to define valuation, we are making the assumption that the metric was fairly valued over the sample studies.
House price to income
This indicator is simply the average house price divided by the average family income in the city. It shows how many years of income would be needed to pay for a house if all the income went to purchase a house. With some manipulations and without changing any conclusions, it also shows how many years it will take for the average household to accumulate enough for the down payment needed for purchasing a house (dividing by 5 for the time to accumulate 20% down payment and by 10 for a 10% down payment).
Based on this measure, it takes 5.1 years for the average Canadian household to pay for a house in Canada. However, there are some important divergences. For example, in Vancouver, it would take 7.7 years and in Toronto 6.6 years, while it would only take 2.4 years in Edmonton, 2.5 years in Winnipeg and 2.7 years in Calgary.
However, as we said previously, while the level of the indicator is interesting, its deviation relative to average is more indicative of valuation issues. This is especially true for Vancouver and Toronto, where the indicator has been higher than other cities for most of the past four decades.
Putting it in a historical context, on average in Canada, it took 2.2 years in the 1980s, 2.5 years in the 1990s, 2.8 years in the 2000s and 3.9 years on average since 2010. These calculations show the scope of the rapid rise in house prices compared to family income over the past two decades.
Using the Z-score to evaluate the amount of overvaluation, the Canadian average makes it significantly overvalued with a score of 3.2. Ottawa at 4.6 and Toronto at 3.1 are also considered significantly overvalued. Montreal at 2.7 is overvalued. Vancouver at 1.4 and Winnipeg at 1.6 are in the stretched territory, while Calgary at 0.6 and Edmonton at 0.6 do not show any signs of overvaluation.
Some may be surprised to see Vancouver as being only in stretched territory rather than overvalued. This is because house prices relative to income have been consistently high over the past decade. The city has had a period of under performance in house prices starting in 2017 due to the foreign buyer tax and higher interest rates, . Nevertheless, the Z-score was still marginally below 2 in 2017.
Mortgage to income
While price-to-income is an interesting measure, it is incomplete and fails to consider a significant economic trend of the past 40 years: the gradual decline in interest rate.
Over the past four decades, the decline in interest rates has led to a gradual but sizable improvement in affordability. This means that households had lower mortgage payment for the same house price or that for the same mortgage payment, they could afford a more expensive house. Moreover, when a household considers buying a house, the size of the monthly payment often matters more than the actual price.
For this reason, the second valuation we consider is the mortgage payment relative to income. This measures the share of income an average family in each city would be spending if they purchase a house at the average price for their city with a 20% down payment, at current mortgage rates and on a 25 years amortization.
Based on this measure, an average Canadian family would spend about 24% of its income on mortgage payment buying an average Canadian house. However, there are some important divergences between cities.
At the upper end of the spectrum, families in Vancouver could expect to spend 36% of their income on mortgage payments. In comparison, a family in Edmonton should expect to use 11% of its income to cover both interest and principal payments.
While the level is indicative of the pressure house prices are putting on household finances, how those pressures compare to historical norms also matter.
Once again, using the Z-score, we find that none of the Canadian cities would be consider overvalued. Only Toronto has a stretch valuation with a score of 1.2. Interestingly, both Calgary and Edmonton, with a score below 0, have better than average affordability.
The difference between the price-to-income and the mortgage payment to income clearly shows that low mortgage rates are the main reason why house prices remain affordable in many markets despite the significant rise in prices in recent years.
Price to rent
The previous indicators look primarily to house affordability. However, households have the choice between owning or renting a home. Hence, if ownership cost becomes disproportionately high compared to the rental cost, some households may decide to rent rather than buy.
We consider houses price relative to rent. This measure can be interpreted as the number of years a household would have to be renting to have spent the same amount of money as owning a house.
It is important to note that the comparison is not perfect because it doesn’t take into account the maintenance cost a homeowner would face and its mortgage payments. Moreover, it also doesn’t account for the fact that owning a house is an asset on the household balance sheet.
Based on this measure, it would take 65 years renting in Vancouver to have spent the same amount as buying a house. In Toronto and Ottawa, it is 59 years and 36 years respectively. At the other end of the spectrum, it would take 19 years in Winnipeg, 22 years in Edmonton, 24 years in Montreal and 28 years in Calgary.
Relative to history, Ottawa stands out with a Z-score of 3.1, suggesting that house prices are significantly overvalued compared to the rental cost. Similarly, Toronto has a score of 2.8 and Montreal is at 2.3. The least overvalued cities are Calgary and Edmonton, both with scores of 0.2. Similar to the price-to-income and for the same reasons, Vancouver is only in the stretched territory with a score of 1.3.
Mortgage to rent
When households consider whether they should buy or rent, an important factor they are likely to consider is how much they would have to spend monthly. Hence, they compare their monthly mortgage payment relative to what the monthly rent would be.
Using the same mortgage payment estimated previously, we look at mortgage payments relative to rental cost. As with the previous measure, it does not take into account the home ownership cost like repair and maintenance, property taxes, etc.
As was the case previously, this allows us to take into account the record low level of interest rates. This measure also shows how big the average mortgage payment in the city is relative to the average rent.
Based on this measure, families in Vancouver should expect, on average, to pay 3 times more in monthly mortgage payments than in rent. This proportion is 2.7 times in Toronto and 2 times in Ottawa and Montreal. In Winnipeg, households would be expected to pay 1.2 times their monthly rent. The ratio is 1.3 and 1.6 of Edmonton and Calgary, respectively.
When looking at the Z-score, it is clear that households in Ottawa, Toronto, and Montreal spend significantly more on mortgage payments than on rent compared to historical norms. All three cities have scores well above 2 at 2.9, 2.4 and 2.4, respectively, meaning that those markets are in overvalued territory.
At the other end of the spectrum with values slightly below 0, the markets of Edmonton (-0.6), Calgary (-0.4) and Winnipeg (-0.1) have mortgage payments relative to rent slightly below their long-term average. Similarly, the measure for Vancouver (0.4) is also only marginally above historical norms, despite having the highest absolute level.
Mortgage interest payment to rent
While households compare their monthly mortgage payments and rent when deciding whether to purchase a house, a mortgage payment is not fully comparable to rent.
As such, the part of the mortgage payment that covers principal repayment could be considered as forced saving and increases the household’s net worth. With this in mind, it could make sense for a household to stretch its mortgage payment relative to rent, especially in a low interest rate environment. With this in mind, we consider the size of the first mortgage interest payment relative to rent.
In level, the mortgage interest payment relative to the monthly rent is the highest in Vancouver (1.7), Toronto (1.5), Ottawa (1.1) and Montreal (1.1). Conversely, it is the lowest in Winnipeg (0.7), Edmonton (0.7) and Calgary (0.9).
Compared to the historical average, none of the cities considered have a Z-score that suggests some potential issues. For example, the measure is well below average in Winnipeg, Calgary, Edmonton and Vancouver, while it is slightly above average in Toronto. However, it is important to stress that this situation is because interest rates are at a record low. This means that despite a very high level of mortgage debt, the interest payment on that debt remains low.
The general conclusion from the various valuation metrics is that, if it were not for the extremely low interest rate, most cities in Canada, especially Toronto, Ottawa, Vancouver and Montreal would be in overvalued territory. It means that the main driver for affordability is the record low level of interest rates. The increased prevalence of remote-working arrangement post-COVID19 pandemic could make such a situation possible, but it is too recent and not yet broadly used to influence the current analysis.
 All the cities covered in our study would represent about 50% of Canada’s population.
 It is important to note that the house prices average in Vancouver and Toronto is above the $1 million and no longer qualify for the mortgage insurance from CMHC.
The views and opinions expressed in this publication are solely and independently those of the author and do not necessarily reflect the views and opinions of any organization or person in any way affiliated with the author including, without limitation, any current or past employers of the author. While reasonable effort was taken to ensure the information and analysis in this publication is accurate, it has been prepared solely for general informational purposes. There are no warranties or representations being provided with respect to the accuracy and completeness of the content in this publication. Nothing in this publication should be construed as providing professional advice on the matters discussed. The author does not assume any liability arising from any form of reliance on this publication.