It Really is About Housing Supply and Canada Needs to Get Building

 

The housing shortage, rising prices and rising rents continue to preoccupy Canadian public policy debates and with good reason.  As of June 2023, median rent for a one bedroom apartment in Vancouver stood at $2,700 and $2,400 in Toronto with rent across Canada up 20 percent over pre-pandemic levels.  Meanwhile, average housing prices in Canada reached $729,044 in May of 2023 – the highest they have been since April of 2022.  Since 2000, residential property prices in Canada have essentially doubled – per capita income have not.  

 

Needless to say, the response in the most Serene Kingdom of Canada has been predictable.  In the name of boosting supply, municipalities starting to chase multiple property owners for tax revenues  (who incidentally are probably renting out the properties they own and helping to alleviate the shortage). Then there is the typical passive-aggressive Canadian story about how seniors are not downsizing and are living in homes with empty bedroom but of course “Policy experts and large city mayors are not suggesting that seniors should rent out their rooms en masse to better use the extra space.”

 

There is indeed a supply issue in Canadian housing, but it is not because there are too many multiple owners who are hoarding empty apartments or existing homeowners who do not want to share their spare rooms.  It is because over the long term the supply of new residential construction has fallen behind the rate of population growth so that housing starts per capita are dramatically lower than they were during the 1970s and 1980s.  Incidentally, this era had even higher interest rates and inflation than today and still managed to keep up with construction.  The accompanying figure plots seasonally quarterly total Canadian residential housing starts (units) as well as the per capita index (with 1961=100) [Data Source: Statistics Canada] for the period 1961Q1 to 2023Q1.  The results are quite startling. 

 

 

 

In the first quarter of 1961, total housing starts in Canada were 34,225 units.  In the first quarter of 2023, they were 55,753 – an increase of 63 percent.  The problem is that in 1961Q1 Canada’s population was 18.1 million while in 2023Q1 it was 39.9 million – an increase of 120 percent.  As a result, when the number of starts per capita are converted into an index (with 1961=100) it becomes quite apparent that despite surging population, we are building fewer new units per person despite a slight upward trend in the total number of units. 

 

The most quarterly housing starts ever were actually  in first quarter 2021 at 73,738 with the average quarterly number of starts in 2021 at 68,612.  In 1973, the average number of quarterly housing starts was 66,883.  Total starts at present are not much different than the peak of the early to mid 1970s.  When you look at the per capita index, the overall trend since 1961 is downward but essentially it appears that after the housing bust of the late 1980s, housing starts per capita have stayed flat at about half of what they were in the 1970s.

 

The baby boom and tail of the boom that entered the workforce in the 1969s to early 1980s was a population surge that was accompanied by new and rising per capita housing construction.  The current surge in population is not being accompanied by rising per capita construction but with construction at the historical per capita rates in place since the 1990s.   That is why housing prices are high. Band-aid solutions that attempt to solve the problems by essentially redistributing existing supply is but another sign of a society that seems to find it increasingly hard to build new things and to get things done.  But then, this is the same society that is dealing with inflation by injecting more money into the demand side of the economy. In the end, government policy to fight inflation is still conflicted with higher interest rates to slow down the economy on one hand and stimulus on the other.  It would be more useful if some of that stimulus went to building housing.

Thunder Bay Employment Flat for Forty Years

My recent Fraser Institute Blog post on employment growth in Canada at the provincial and CMA level since 2007 appears to have attracted a fair amount of interest if only based on the hits via my Linkedin page.  The article was posted on the Fraser Blog on December 4^th and by December 14^th, it had garnered 1,515 views.  The interest has been quite pronounced from Linkedin profiles in Ontario and of course particularly from the Thunder Bay area. As a follow-up, I decided to look at employment levels in Thunder Bay and Greater Sudbury from a longer-term perspective using data from Statistics Canada.

Now Statistics Canada has annual province level unemployment rates and employment data available on its site from 1976.  Its annual CMA level data only appears to go back to 1987.  So, in order to generate CMA employment levels and unemployment rates for Thunder Bay and Greater Sudbury prior to 1987, what I did (acting on the suggestion of my Lakehead colleague Rob Petrunia) was run regressions of CMA level employment and unemployment rates for both cities on the Ontario data along with a time trend variable.  The assumption is that employment levels and unemployment rates in the two cities should reflect what is going on in the province as a whole. The regression results were then used to estimate fitted values for Thunder Bay for the period 1976 to 1987 and for Sudbury from 1976 to 1990 (Sudbury data starts in 1990). 

The results are intriguing.  Figure 1 plots the unemployment rates in the two cities from 1976 to 2016 and there seems to be some good news here.  While unemployment rates in both cities fluctuate a great deal over time, they have generally trended downwards since the late 1970s.  The average unemployment rate in Thunder Bay between 1976 and 1985 was 9.7 percent while in Sudbury it was 11 percent.  Over the period 2010 to 2016, Thunder Bay’s unemployment rate was 6.1 percent while over the same period in Sudbury it was 7.5 percent. 

 

However, the good news seems to end when employment levels are examined in Figure 2 – at least for Thunder Bay.  Sudbury has seen its employment grow over time while Thunder Bay has essentially remained flat. In 1976, estimated total employment (full and part time) in Thunder Bay was 61,224 and in Sudbury it was 60,475.  By 2016, Thunder Bay’s employment was 60,100 while in Sudbury it was 81,700.  In other words, over 40 years Thunder Bay has essentially remained flat in terms of its employment level – indeed there has been a slight decline of 2 percent since 1976.  As for Sudbury, its employment level has grown by 36 percent since its estimated 1976 value.

 

A declining unemployment rate when total employment is growing can be seen as good news.  A declining unemployment rate when total employment is declining means that your labour force is actually shrinking faster than your employment level.  For Sudbury, a lower unemployment rate is good news given that it has been accompanied by rising employment.  For Thunder Bay, a declining unemployment rate is a misleading indicator and masks the moribund nature of its economy given that its employment level has been essentially the same for 40 years.