I recently highlighted the fact that aggregate Canadian house prices appear to be significantly overvalued relative to GDP per capita (one of the key determinents of income….all things being equal, higher income levels should equate to higher house prices).
Here are the key charts from that post:
As several of my readers have noted, an examination of house prices in Canada at the national level may have implications for the broader economy, but it does little to reveal local variations in house prices.
This is certainly true. I have long held that our national housing bubble is inflated primarily because of a handful of markets and specific market segments. Nevertheless, I’ve been hesitant to predict that too many locales will escape unscathed given the widespread element of mass psychology. It’s hard to find a part of the country where housing is looked upon as anything other than a fantastic investment and a great store of value. Furthermore, the widespread credit excess caused by the loosening of CMHC mortgage insurance standards knows no Canadian boundaries.
So with that in mind I undertook the significant challenge of analyzing provincial house prices based on one primary determinant of fundamental value: GDP per capita. This was done by accessing provincial resale data from CREA, and population and GDP data from Stats Canada. Today we will examine BC east to Ontario. Tomorrow we will look at the rest of the eastern province with the exception of Quebec (apologies to my Quebec readers as their house price data is not recorded by CREA….I’ll see if I can get my hands on it, but it won’t be this weekend).
I’ll let the charts do the talking for BC:
I’m not sure that much else needs to be said about these findings. Not too many people would be surprised to learn that house prices in BC are exceptionally out of touch with economic fundamentals. Let’s move on to a finding that quite surprised me…
You may recall that Robert Shiller, arguably the world’s leading expert on real estate bubbles specifically pointed to Alberta as a bubble zone. House prices have in fact risen substantially, but so too has GDP per capita.
Now what I want you to note is that in this case house prices have apparently lagged GDP growth for nearly 25 years, only to catch up in 2006. Let me suggest that this is quite likely a reflection of the starting point. Throughout long enough periods of time, we should expect gains in house prices to pace gains in GDP. When it doesn’t, it is far more likely a reflection of a skewed starting point. From 1981 to 1986, house prices lost 22% of their value as a pre-existing bubble popped. If in fact the 1981-1986 housing bust was the result of a pre-existing bubble and simply resulted in house prices realigning with GDP growth, it could be argued (at the risk of being accused of data mining) that a different starting point would be a better indicator of current over or under-valuation.
With that in mind, I advanced the starting point by 5 years, aligning it with the bottom in the housing bubble and with the bottom in the drop in per capita GDP as seen on the chart.
You can see that from this point, GDP and house prices track each other much more closely…..until the mid 2000s that it. I will leave it up to my readers to decide which chart tells the more compelling story.
Interestingly, when the average house price is plotted against average GDP per capita, it seems to confirm that there is a level of overvaluation based on historic norms, but perhaps not as significant as the second chart above.
Of all the Western provinces, Saskatchewan seems to be the most fairly valued, perhaps in part due to restrictions on foreign ownership of farmland until the regulations were loosened in the past few years. Yet once again we seem to have the same problem of house prices apparently underperforming GDP for 20 years. What is interesting is that Saskatchewan never experienced a significant boom and bust. Could it be that houses and land in Saskatchewan have managed to avoid the sort of “easy riches” mentality that is the hallmark of an asset bubble?
Even if we advance the data to look for a line of better fit, we note that by all accounts, Saskatchewan seems much more fairly valued. The best line of fit for the two data series begins in 1990 and then deviates markedly about the same time as other parts of the country, namely the mid 2000s.
Yet if we divide Saskatchewan house prices by GDP per capita, we note the following:
There may in fact be a small amount of froth in some parts of the Saskatchewan market, but by and large, it is much more fairly valued than other areas.
Manitoba is surprisingly bubbly….
Note that house prices significantly underperformed GDP for much of the 90s and yet there appears to be significant overvaluation at present. Advancing the data set to find a cleaner fit would only make the housing overvaluation that much more acute.
When house prices are divided by per capita GDP, the variation from the mean is particularly noticable.
Once again, the bubbly level of house prices in Canada’s most populace province is hard to miss…
You’ll note the ‘bubble’ (centred over 1989) which led to a fairly painful housing correction in the early 90s followed by a decade of stagnant growth. Note the relative variance of that blip over current price levels.
It is perhaps even easier to spot when house prices are divided by per capita GDP.
Indeed it appears that if there is an east/west divide that separates the ‘bubble’ provinces from those more closely in line with their fundamentals, that line appears to be somewhere east of Ontario.
We’ll look at those eastern provinces tomorrow to try to get more of an answer…
Cheers for now,