Become a member. Sign up here

Become a member

Already a member?

Forgot your password?

Redux: House prices and income growth in Canada...bigger and better

JULY 07, 2011

Readers of this blog will recall that a few months ago, I did a two-part series on house prices and income growth in Canada.  My philosophy is that when better data becomes available, we analyze it.  In this case, that first series used a data set from Stats Canada that unfortunately only extended back to the early 90s.

However, I have been able to access a great data set from the Conference Board of Canada:  Personal disposable income.  This data set actually goes back further than the house price data set.  Personal disposable income is a great measure since it levels the playing field in terms of differential tax rates across the country.

Stats Canada defines personal disposable income as:

“the amount left over after payment of personal direct taxes, including income taxes, contributions to social insurance plans (such as the Canada Pension Plan contributions and Employment Insurance premiums) and other fees. It is a measure of the funds available for personal expenditure on goods and services and personal saving for investments as well as personal transfers to other sectors of the economy.”

As such, I am quite pleased to present the latest on house prices and incomes in Canada.  Similar to the house price and GDP series, there are a couple notes about the data:

  • The data set for Quebec is from the Fédération des chambres immobilières du Québec.  In 2002, they switched to a weighted average, meaning data before and after the switch are technically not comparable.   I've done it anyway as I believe the difference is negligible.
  • In the early 80s, there was a significant house price correction in the prairie provinces which heavily skewed the charts showing the cumulative growth in house prices and incomes.  I instead chose to begin these data sets where the house price/personal disposable income ratio again reached its long-term norm.  The data purists may contend that I have manipulated these data sets since the start date is advanced by a couple years.  I'll take my chances.

 

Here's the data.  Discussion to follow...

Canada

house prices incomes canada

house price income ratio canada

 

BC

house prices income bc british columbia

house price income ratio bc british columbia

 

Alberta

house prices income alberta

house price income ratio alberta

 

Saskatchewan

house price income growth saskatchewan

house price income ratio saskatchewan

 

Manitoba

house price income growth manitoba

house price income ratio manitoba

 

Ontario

house price income growth ontario

house price income ratio ontario

 

Quebec

house price income growth quebec

house price income ratio quebec

Apologies to my east-coast readers.  I'll try to tackle your data later this week.

 

Discussion:

Certainly we can see that house prices have outpaced income growth over the last decade in all provinces.  The ratio of house prices to personal disposable income is also at all time highs in all provinces except Alberta.  I'll go ahead and predict exactly what the detractors will say:  "You have to adjust for purchasing power as interest rates have fallen."

Fair enough.  I don't at all dismiss the notion that interest rates have been a major driver of house price appreciation.  The general relationship is quite obvious.

But I do think they take a back seat to the general availability of credit and the willingness of Canadian home buyers to take on said credit.  I maintain that this is the real driver of house prices in Canada over the past decade:

 

But for those who insist that it is entirely interest rate-driven, perhaps they could explain how we had a positive correlation between interest rates and house prices for a four-year period during the past decade.  I'd love to hear that explanation.

house prices interest rates

 

While falling interest rates have enabled people to carry larger mortgages and have kept affordability seemingly within its long-term range, I would heavily caution against relying on such affordability measures as a barometer of housing market health.

Finally, the impact that BC is having on the national average can certainly be seen in the following chart depicting the price/personal disposable income ratio of all provinces:

Note that Ontario is right on the Canadian average, BC is way above it, and all other provinces fall below it.  This is not to say that there is not some froth and unsustainable house price increases in other provinces, rather it should highlight just how far BC prices will fall.

For those interested, I will be providing historic price/personal disposable income data for the following cities:

Victoria, Edmonton, Regina, Saskatoon, Winnipeg, Hamilton-Burlington, Kingston, Kitchener-Waterloo, London, Ottawa, St. Catharines, Thunder Bay, Durham region, Windsor-Essex, Saint John, and Halifax. 

Watch for that later this weekend.

Cheers,

Ben

Posted in:

Ben Rabidoux
By Ben Rabidoux

Enjoyed this Post?

Subscribe to our RSS Feed, Follow us on Twitter, Subscribe by email or simply recommend us to friends and colleagues!

16 Comments

  • Etienne said:
    • 2 years, 9 months

    can you get disposable income for Montreal too (and associated house price)?

    Reply
    Post a comment
  • Ben Rabidoux said:
    • 2 years, 9 months

    I probably can. Waiting to hear back from FCIQ.

    Reply
    Post a comment
  • Statsman5000 said:
    • 2 years, 9 months

    Hi Ben

    If possible could you calculate the r-squared values of house prices vs mortgage debt as well as house prices vs interest rates? I would say the variables in both relationships are highly dependent on each other.

    The mortgage debt vs house price high correlation argument is an obvious one because essentially they are the same thing. It's obvious these variables have a positive relationship - as house prices rise, so will mortgage debt. What's important to get out of this is - these people are at greater risk

    In my opinion, the house price vs interest rate correlation is a more impressive argument because it's more of an external factor. For correlation, I agree that there are other factors than interest rates that affect the house price vs interest rates correlation. Maybe, demand is cooling because a lot of the new homebuyers have already taken advantage of the low rates, etc (I am not an economist)... You can't expect such a strong correlation compared to mortgage debt vs house prices. BUT, the negative correlation is also obvious and is not a guarantee.

    Why I ultimately think interest rates is a more importatnt correlation is because it can be used for predictions. For example - In the future, if you know interest rates are going up - you can conclude there will be a % chance that house prices drop. In the future, you cannot say "I know debt will increase", so that means house prices are also going to increase.

    Those are my 2 cents.

    But, as always, love the insights. Looking forward to your next post and book...

    Reply
    Post a comment
  • Ben Rabidoux said:
    • 2 years, 9 months

    "The mortgage debt vs house price high correlation argument is an obvious one because essentially they are the same thing. It's obvious these variables have a positive relationship - as house prices rise, so will mortgage debt."

    Nominal mortgage debt, yes. But why would it rise against GDP? That is the variable I focus on. Rising house prices and rising mortgage debt relative to GDP are not the same thing. I see it as a bad sign.

    I'm going to do a post next week addressing some of your other points. Thanks for the input.

    Reply
    Post a comment
  • Statsman5000 said:
    • 2 years, 9 months

    Just another idea WAY out there -

    What I think would be amazing is if you could create a multivariate predictive equation which could calculate the probability/degree of house price decline/increase using logistic regression. Variables can be things like demographic, external markets, internal markets, interest rates.. and the list goes on. The regression coefficients essentially scale the effect that that variable has on house prices in a country. When you plugin in variable values (and relate to previous dates), the model will determine whether or not house prices will increase/decrease and by how much. Believe it or not, but you can model stuff like this. A lot of the information you already have is in your previous posts.

    I'd suggest using an application like SPSS. Would be a good idea for a thesis. You probably couldn't fit it in as a post on your site. It may have been attempted in the past. Don't really know. Have you seen a thesis on this before?

    Anyways thanks for your response. I am not much of an economist so please make a post (or reply to this) about what you mean by "relative to GDP" and "nominal debt" and "nominal interest rates". One thing that I've liked about this site is how well you can explain especially with the primers. Must be the instructor in you. Cheers.

    Statsman5000

    Reply
    Post a comment
  • aaronH said:
    • 2 years, 9 months

    Excellent post Ben!! By far this is he best blog that I've come across on the housing situation in Canada. Kudos.

    Statsman, see Wiki for your definitions. Perhaps you could construct such a model?

    Reply
    Post a comment
  • Statsman5000 said:
    • 2 years, 9 months

    Ben has a lot more knowledge than me for what's going on in the real estate markets. He also has a ton of statistics. I wouldn't be able to recognize all the parameters in a RE price predictor model. Just an interesting idea that came to my head. Would be a lot of work too

    Reply
    Post a comment
  • Etienne said:
    • 2 years, 9 months

    off topic, but you you have a subscription to Capital Economics (seems it's 6000GPB per year!)
    They have interesting articles a free trial would allow to read like:
    http://www.capitaleconomics.com/canada-economics/canada-economics-weekly...
    and
    http://www.capitaleconomics.com/canada-economics/canada-economics-update...

    Reply
    Post a comment
  • jwg said:
    • 2 years, 9 months

    BC, having in the more populated areas, the best (as many people see it) weather in Canada, attracts retirees from other parts of Canada. These people will likely have on average higher savings and declining incomes. Might that explain some of the BC divergence from the rest of Canada?

    Reply
    Post a comment
  • Ben Rabidoux said:
    • 2 years, 9 months

    It's an interesting thought. It have some impact. We could look at net provincial migration to get an idea of how many people are coming to BC. I think it will be entirely insufficient to explain the massive divergence between house prices and incomes. And regardless, even if it were true, to whom will these people sell at these prices?

    Reply
    Post a comment
  • Chris said:
    • 2 years, 9 months

    BC had a net provincial *emigration* last year. (Yes, even with the incoming retirees and Asian immigrants, more people left BC than arrived.) I don't think this changes your thesis though -- it might even support it.

    Reply
    Post a comment
  • Financial Newbie said:
    • 2 years, 9 months

    Ben,

    Something I'm not sure if you've touched on - falling wages.

    I have two home owning friends in the periphery of the GTA (think further than Milton). They're feeling more comfortable with their highly mortgaged homes (and owning highly-leveraged rental properties as well) because, while they agree places like Toronto and Vancouver are overpriced, they think their smaller town has dodged the worst of the housing mania that's driven up housing prices so drastically.

    To some degree I have to agree with them on this, but I still caution them that they may not be out of the woods yet for one very important reason - falling wages.

    If we find ourselves in a moderate-to-major deflationary spiral (which I think we will - anyone check out the NFP numbers today? Yikes...), doesn't history teach us that many of the jobs that disappear first are in the periphery of major cities, while the major cities themselves become "last bastions" of employment?

    I guess what I'm getting at is that, even if Toronto and Vancouver are so obviously overpriced, if the major urban centres offer a majority of the well paid jobs as the Canada slips into depression, wouldn't that help to prevent housing (and rent) prices from falling as drastically as we might think? And perhaps the suburbs and places where there is not industry/employment, while not as overpriced, could potentially lose more value in real terms?

    I base all of these ideas on the concept that housing prices & rents are a factor of average income within that geographic area. Don't know if it works in relation to developed countries, but certainly seems to apply to what I've seen in my travels through Mexico, Thailand, Cambodia, Vietnam, Indonesia, etc. (though in that case, it could also be a factor of lack of infrastructure in the periphery of major urban centres and in rural areas).

    Your thoughts on this would be appreciated.

    Cheers,
    -FN

    Reply
    Post a comment
  • Dirk said:
    • 2 years, 9 months

    We know that in Saskatchewan, after 3 positive years, the last 3 months people are again leaving Saskatchewan for interprovincial migration. Homes in Saskatchewan are now as much, or more, than most of Alberta (ie bang for buck, Saskatoon is more expensive than Edmonton). Saskatchewan now relies on a booming first nations birth rate, and third world migration for its very moderate population growth.

    At least Alberta has high wages (and low taxes) which make expensive housing at least doable, if not desirable.

    For the REST of the country, this is NOT sustainable.

    We know that Canadians now hold MORE debt per person than the USA at the height of its ridiculous credit crunch...

    We know that IF any major sector laid offer workers, that FEW (if any) could afford even the mortgage on their homes

    Savings are stagnant
    The Federal CONS are telling us to rejoice because credit is growing at a SLOWER rate (but still growing)

    The government should slowly (so as to not insite a real estate crash) raise the MINIMUM down payment to 10% of the value of a home. If you cannot afford 10%, you are looking too high end, OR need to save for a few more years. This will hold house prices (but not drop them) so that in 5 to 10 years, incomes and inflation can finally catch up.

    In Saskatoon in early 2007 - the AVERAGE house/condo/trailer/duplex cost $150,000
    In Saskatoon in mid 2008 - the AVERAGE house/condo/trailer/duplex cost $330,000
    Incomes went up about 4% that year...

    Now how does that make sense?
    How is that sustainable?

    And to those who think that's a "deal"?
    An actual newish bungalow (within last 20 years) that's not in one of the worst areas in Canada... at LEAST $400,000 (if not $500,000) in Saskatoon

    And the Median Family Income in Saskatoon is more than $10,000 LESS (each year) than Alberta....

    Reply
    Post a comment
  • Mark Chang said:
    • 2 years, 9 months

    Hi Ben, I don't agree with BC ratio at 18. A lot of house in Lower Mainland has a secondary suite getting $600 rent alone; which would support $100,000 worth of house. With 2 person making minium wage in 2012 at $10.25 per hour would have a combined disposible income of $37,515. Looking at house for sale in North Delta with a secondary suite for $480,000 would give a ratio of 10; not 18. Anyone looking at a similiar house in Burnaby cost $900,000, and even more costly for Richmond, Vancouver, etc. Yes, ratio would be much higher even with one or two suites; so you would need to make twice the minium wage, not hard for most people working.

    Reply
    Post a comment
  • Ben Rabidoux said:
    • 2 years, 9 months

    I appreciate the insight, Mark, but I think you're reaching a bit too much. If you can find data showing just how prevalent this phenomenon really is, I'd love to have a look at it. Short of that, we have to acknowledge that the personal disposable income data is reliable and factual, while what you are suggesting is purely anecdotal. Until I see solid stats to support your assertion, I know where I'd put my faith...

    Reply
    Post a comment
  • pm said:
    • 1 year, 11 months

    Very nice analysis. Look forward to the Atlantic data.

    Reply
    Post a comment
Post a comment