I’m currently working with some data to show the historic relationship between house prices in Canada and various other factors such as inflation, population growth, immigration, income, mortgage debt, rents, etc. I’ll share the data as I complete it and then compile it into one giant post that will become the next primer.
Today I’ll share a couple of the graphs I’ve compiled examining house prices and per capita GDP. The house price data is from CREA resales and must be ordered from them at a cost. To calculate GDP per capita, I used data from the World Bank (GDP) and Stats Canada (population).
I’ll let the first two graphs do all the talking…
For these last two charts, I was curious what the correlation would be between year over year (Y/Y) changes in house prices and the year over year changes in per capita GDP.
What I want you to notice is that the two data series have a correlation of 0.366 or moderate. However, if we subdivide the data series into the periods 1980-2000 and 2001-2010, we find that it tells two very different stories.
The data sets for the 1980-2000 period highlighted in yellow were more strongly correlated with a coefficient of 0.5312.
Yet the data set for the 2001-2010 period highlighted in green was weakly correlated with a coefficient of 0.1502.
My position has been that our current credit bubble had its origins in the aftermath of the dot-com bubble and the associated tanking of interest rates. At that time it was nothing but a mild overvaluation. But gas was thrown on the fire between 2004 and 2009 as the loosening of mortgage lending standards and a new prevailing mentality towards housing took hold. The combination of shifting mass psychology and cheap, readily available credit has been the fuel for the made-in-Canada housing bubble.
More to follow…
Cheers,
Ben
Financial Insights 
I’m with you on most of your stuff, but I think the last two graphs (Y/Y GDP change vs. Y/Y House Price change) show that you’re suffering from a bit of confirmation bias. 0.5 is not that strong of a correlation and while you COULD interpret them as lending strength to your arguments I find them mostly inconclusive. It actually hurts your argument when you try and force inconclusive data to support your views. Your first two graphs more clearly illustrate the change in GDP per capita to house price.
Thanks for the honest response. Strength of correlation is interpreted differently by different people. Most suggest that 0.5-1 is a fairly strong correlation. The point is that the correlation is much stronger during the 1980-2000 period and virtually non-existant during the 2000-2010 period.
You would have to perform a test of significant to determine if correlation is significant or not. 0.50 can be significantly different from 0 with a 99% accuracy rate for large sample sizes, or we could conclude that the correlation is no different from 0 for a 0.5 correlation (or even higher) for smaller sample sizes…. In order to confirm this information, these correlations would have to be tested before we can conclude if they are significant or not.
Yes, .5 isn’t much of a correlation, but I wonder what it would be if we could separate the two solitudes. Again, I would like to see this broken into western and eastern Canada.
The first two graphs certainly show that you are on to something, but we need to bring it into focus? I won’t be surprised if the data you need isn’t available.
I have seen speculation on other blogs that this election is going to be west against east with a strong western separatist bias. I’m not convinced of that, but the economy of the west is certainly not the economy of the east. Never was, really.
I’ll have a look at the Teranet data and reproduce the same chart. Unfortunately it only looks at larger centres and the data does not extend as far back, but we should still get to see some trends.
This blog continues to be simply outstanding on regular basis. I almost feel guilty that I get to read it for free. A sincere thank you Ben for your efforts and analysis.
My 2 cents for the day… from a Calgary perspective… a big part of me looks at Vancouver and can’t help but wonder, “why can’t Calgary’s price to income ratio soar to Vancouver levels before the inevitable return to its long-term trendline?” Since I have no answer for that question, I conclude reasonably that it very well could happen, especially since rates are not going up any time soon. There’s a big issue that people like me are trying so hard to grasp… the timeline. I think “greater fools” in places like Calgary are actually going to do surprisingly well over the next 3 to 5 years, even though the past 3 have been a sideways grind. And renters will end up being bitter (again) because they only attempted to educate themselves, without realizing the house contrarians were too early with their forecasts. So they missed the boat. Again. Meanwhile the fools who waste time watching American Idol end up doing just fine with their real estate decisions… all because they kept watching Global at 11.
This is a really outstanding blog.
Brad Calgary prices could rise in the future but don’t think it will happen and if it does it won’t be anything near Vancouver levels. Calgary boomed because of three main reasons one was the relaxation of the mortgage rules/interest rates falling the second was natural gas prices and the third was a under supply of housing. These caused housing prices to spike that then gave people the buy now or be priced out forever mentality and prices only can go up idea. Now mortgages are being tightened, gas prices are in the tank due to new supply and technology (oil shale/fracking and horizontal drilling) and lots of new houses and condo’s are being built around the city.
They say Calgary is a oil and gas city but it is the natural gas that really powers and pays the bills in Alberta. Unless the price for natural gas spikes housing prices really can’t spike again in the city. However, that doesn’t mean house prices will fall much or quickly. Housing will likely be sticky on the way down unless something really pushes housing to fall.
Western democracies had to replace lost manufacturing jobs with construction jobs. The gig is almost up.
Brad,
I think you hit the nail on the head with that one. A wise friend of mine once told me “The market can remain irrational longer than you can remain solvent”. In other words, don’t bet the farm on an irrational situation correcting itself in the near-term. As long as jobs remain strong in a city people can keep drinking the real estate kool-aid and feeding the suppy of “Greater Fools”. I’m not going to bank on it, but it wouldn’t surprise me. Pyramid schemes and bubbles work just fine if you’re not the last one in.
“Again, I would like to see this broken into western and eastern Canada.”
Well put, prior to the post-2000 global bubble regional house prices in Canada were weakly correlated, which means using national house prices versus GDP does not tell you a lot in the short term. In particular house prices in BC went up substantially during the late 80’s-early 90’s Toronto bust.
My feeling is if you used Ontario house prices versus GDP you’d see a much stronger correlation.
Referring to your first graph, in the period 1984 to 1989 you show about a 70% increase in house prices. I don’t remember a reality anything like that, at least not in Ottawa where prices were pretty stagnant.
Still, in the five years 1984 to 1989 you show a 70% increase followed by a 3% decrease, followed by several more years of increases. A 3% correction wasn’t too bad after such a large run up. Was the run up a bubble?
For the seven years 2001 to 2007 you show only a 65% increase. Not as bubbly, and followed by only a 1% correction.
How do we reconcile the bubbly period in the 80’s with our recent run up? Was the run up somehow justified by the fundamentals then, but not now? What role did inflation or interest rates play then and now?
Finally, I’m not sure your mathematical relationship is correct. If the YoY change in house prices had lagged YoY change in GDP for the past 40 years, by say 5%/year, and then suddenly lead by 25% for just one year, would house prices be 25% over valued or still horribly undervalued over the period? I think you may need to examine the second derivative over the period. In other words, is the sum of the under and over during the period positive or negative?
“If the YoY change in house prices had lagged YoY change in GDP for the past 40 years”
It would be interesting to verify if there is a statistically significant lag between the two time series.
To me, growth in house prices should lead GDP growth, not the other way around.
“To me, growth in house prices should lead GDP growth, not the other way around.”
I think it depends on whether RE activity (construction/sales) is the main driver of the GDP growth. In a healthy economy, GDP growth generated by manufacturing/tech/trade/mining etc. should lead the increase in house prices via expansion of the work force and rising wages.
‘If the YoY change in house prices had lagged YoY change in GDP for the past 40 years, by say 5%/year, and then suddenly lead by 25% for just one year, would house prices be 25% over valued or still horribly undervalued over the period?’
They would still be horribly undervalued for the period if using the methodology I used in my second graph. This should all come out in the wash.
”How do we reconcile the bubbly period in the 80′s with our recent run up? Was the run up somehow justified by the fundamentals then, but not now? ”
At the tail end of the 80s ‘bubble’, house prices were only 15% above the increase in per capita GDP over the previous decade. That led to a decade of stagnant house prices. Since 2003, house prices have outpaced per capita GDP by 27%.
Hi Ben,
Thank you for the answers, but I really hope you can find the time to crunch some more numbers. If we are playing a game of catch up then we don’t actually have a situation of over valuation. The big problem we have is, what is fair value and when was there fair value. What is our base reference point? Were houses at fair value in 2001? Some of the geysers on this blog may say, “Humbug?”
There is some bubbly stuff going on, but there was bubbly stuff going on in the 80s. What makes the present bubbly stuff systemically significant?
As regular readers know, I visited my parents in southern California early in March. They have spent their entire lives in banking and finance. My father is invalid and doesn’t say much, but my mother volunteered that the amazing decent in the US dollar doesn’t affect their life style much. They like to take $40,000 cruises in July and August because the desert is hot, hot, hot!!! Her point was that the cruises are priced in US dollars. Heck, damned near everything is priced in US dollars, so it doesn’t matter to Americans what the stupid thing is worth.
Not the same situation for us Canadians. Our Loonie has appreciated from 63 cents in 2003 to $1.02 now against the US dollar. Imagine, if you were retired and thinking about taking that same $40,000 cruise if our Loonie plunged back to 63 cents.
Systemic risk for Canadians, or Irish, or Portugal, or Spain, or what ever, is a very different story for us than it is for Americans. The products and services we consume, or even sell such as oil and potash, are priced in US dollars.
From my point of view, the very fact that so much of trade is priced in US dollars gives us a kind of “Get out of jail free” card. The US $ is going to continue its relentless plunge. It gives us a lot of room to misbehave. We may not deserve it, it may not be right, but it remains the situation at hand.
This is a simple matter of fact. Canada’s star is rising, the American’s star is setting. In the interim, we have “interesting times.”
I’ve just looked at the raw Teranet data. I’m not sure if it’s worth crunching at this point. Data for Ontario cities only go back to 1998. I’ve emailed OREA asking for data. In the meantime, it would also make sense to look at GDP/capita in each province. That would certainly be an interesting graph and would be much more telling. I’ll see what I can do.
I’m not so sure that fair value is as opaque a concept as you suggest, John. If we use the 80s boom as an example, you’ll note in the first graph that the peak of those boom years resulted in a house price/per capita GDP ratio of 6. We’re at 7.2 today.
We know that by every measure of price/income and price/rent we are in unprecedented territory. Certainly the bubble is more pronounced in the west, but I’m not so sure that we haven’t developed some significant froth over this way as well.
I’ll keep looking into it.
“If we are playing a game of catch up then we don’t actually have a situation of over valuation”
I think Ben’s price to rent and price to income indicators provide a decent gauge of valuations. This gauge, as like P/E ratio, tends to be within the bounds of long-term price fluctuations. What’s important to understand is that there haven’t been many times in recent history where the price-rent indicator has been at “fair” value. That does not mean the indicator is extraneous, only that factors pushing housing from fair value, such as increasing home ownership rates, relatively loose mortgage lending conditions, and a general perception of safety around tangible assets, are commonplace these days. None of these excuse a permanent deviation from the underlying earnings fundamentals.
We have examples of price-earnings reversions in pretty much every past market bubble. I can’t yet think of one that hasn’t subsequently deflated.
John,
Fair value is whatever the buyer considers fair/reasonable and is willing and able to pay. Thus, it is totally subjective.
If you mean ‘sustainable price’ for a specific commodity in a specific market, that’s totally different and certainly more relevant in the context of regional housing markets in Canada.
To me, sustainable pricing is influenced by: 1. Income trend and income stability; 2. Credit availability and cost; 3. Debt servicing ability throughout the duration of ownership, which, in turn, is related to income and credit cost; 4. The ability to cover potential losses at the time of selling (i.e. negative equity), either with savings or with access to further serviceable credit.
Price/Income, Price/Rent, Equity/Price are gauges reflecting the above. Credit availability and cost is the wild card behind all the madness we area seeing in the world today.
Re. the Canadian Star, I’m not as optimistic. Despite all their problems, countries like the US and Japan are still forces to be reckoned with when it comes to innovation. The tide can turn (no pun intended in Japan’s case) given the right leadership and sentiment change in the people. We as a nation are truly lacking in the innovation department, for whatever reason. Sad to say, but the term ‘sclerotic’ always comes to mind whenever I think about Canada.
@Ben,
I’m a bit uneasy about trying to find a correlation between house prices and GDP, as the two can be confounding to each other, especially in the case of a housing boom where RE related activity may constitute a disproportionate contribution to GDP relative to the rest of the economy. A bit too ‘chicken and egg’ to be useful, in my opinion.
@ATP: Canada has its fair share of innovation over time. We needn’t be ashamed of our contributions to the world.
As for being sclerotic, I think that is what is working in our favor at this point in time. It is sort of a tortoise and the hare situation.
We can congratulate ourselves on our foresight and wonderful planning in avoiding the financial contagion that overtook the world over the past few years, but the reality is it was our relative isolation that saved us. Nobody invited us to the party.
It is Spring and -10C this morning. There have to be some advantages to freezing to death eight months of the year.
@John
“We can congratulate ourselves on our foresight and wonderful planning in avoiding the financial contagion that overtook the world over the past few years, but the reality is it was our relative isolation that saved us. Nobody invited us to the party.”
Isn’t there a possibility that the differences in the Canadian banking/housing/financial systems vs. the rest of the world have actually deferred the financial contagion rather than led to us avoiding it? I find it hard to believe that, somehow, Canadians had so much foresight/willpower/resistance to Kryptonite that we avoided economic catastrophe while some of the greatest nations in the world fell victim to the same trap.
I suppose I’ll know if my fear is valid if/when the day of Canadian reckoning comes.
Pingback: House prices: Examining inflation and supply/demand factors | Financial Insights
Pingback: A regional look at Canada’s housing bubble: Part 1 | Financial Insights