One swallow doesn’t make a summer. Seasonality of Real Estate in Vancouver

Spring is the season when flowers bloom, birds migrate, and people buy and sell real estate, or so people think. Does the data back this up? Let’s see.

Seasonality in inventory and sales

Once again, the data was downloaded from Steve Saretsky’s website in October 2018, and manually updated since then. They correspond to the city of Vancouver, not the total REBGV region.

The following graphs show the typical plot of sales vs inventory, per unit type:

Right away we can see that indeed there are yearly cycles in the sales and inventory, but interestingly, it also appears that the numbers are anti-correlated. To inspect this, let’s take the 6 month averages (to smoothen the data) and plot sales vs inventory now on different axes:

Now the anti-correlation is more apparent. I put a purple arrow to highlight the periods where the anticorrelation is more pronounced.

The following plots show the actual correlation plots:

This shows that supply and demand are not completely decoupled. And this makes sense, as both of them are not completely elastic. When there is a period of very high demand (2015-2017), supply can’t keep up in the short run, and hence inventory drops. When there is a period of low demand, supply increases.

SAL per unit type

In our previous blog post, we determined that the ratio of Sales to Active Listings (SAL) is a good predictor of future price changes, so now let’s examine how SAL changes each month of a given year.

One way to visualize the SAL is to look at a heat map, where high numbers are coloured in red, and low numbers in blue. The heat map of the SAL per unit type is shown below:

A few observations are apparent from these tables:

  • On average, it is true that March-June tend to have the highest Sales to Active Listing (SAL) ratios.
  • However, the Coefficient of Variation, which is a measure of how much a number changes with respect to the average, is much greater for a given month across different years, than for a given month compared to other months of the same year. This is summarized in this table:
Unit typeAverage CoV within a yearAverage CoV of any given month across different years
  • What this means is that in a ‘hot’ year, winter is likely not coming, Dec and Jan, traditionally the slowest months, will still be busy. And the same applies the other way around. In ‘cold’ years, spring is likely not coming. As 2018 turned into a ‘cold’ year for detached houses, and started to slow for strata units, many people think that things will magically pick up in spring of 2019. The data does not support this proposition. March 2019 is more likely going to correlate with February 2019 than with March 2018.
  • In other words, the SAL in upcoming months is more likely going to correlate with the SAL of previous months than of the same months from other years. This actually makes sense. Factors such as credit availability, interest rates, and speculation waves do not distinguish ‘spring’ vs ‘winter’.
  • Another way to represent this observation is to look at the average SAL per year vs average SAL of a month across years:
  • The error bars are the standard deviation, and they are generally larger in the first plot than in the second (since in both cases we have 12 data points, the standard deviations of the two can be compared). Note that for 2019 we only have data for two months so no error bars are displayed.
  • Note that the error bars are smaller in the second graph, supporting the point we made above: SAL correlates better with other months of the same year than with the same month of different years.

In the long run supply is elastic

Although we showed above that supply is inelastic in the short run, and periods of high demand result in a drop in listings and increases in prices, people often don’t realize that in the long run, supply IS elastic. The following graph shows a typical real estate cycle showing how construction often booms after a boom. This was first observed over 100 years ago by economist Henry George. Unfortunately as construction is slow and takes years, it ends up delivering the new supply after the demand has decreased, exacerbating the cycle:

Indeed, if we look at the data for units under construction in Vancouver, you can see how there will be a large increase in supply of new units in the near future, particularly for condos (‘multiples’, source):

Indeed, it is estimated that 40,000 new units will hit the market in the next two years. To put this into context, today the total inventory at the MLS is about 13,000 units. An increase of 20,000 units per year would be very significant. Josh Gordon, assistant professor at the School of Public Policy at Simon Fraser University said: “In bubbles, people underestimate the elasticity of supply,” he says. Higher density housing can substitute for detached homes, for example. “Eventually, supply will catch up to that surge of demand. When that happens, you can have downward pressure on prices.”


We have already discussed in the last post that SAL dictates price changes. Looking at the data for the city of Vancouver, it is true that spring is the season with the highest number of sales to active listing ratio on average. However, upon closer inspection, we realize that there are stronger forces to the SAL than the 12-month cycle. That is, SAL is more likely going to correlate with trends of the recent past than with the same month from different years. Part of this is because in the short term, the number of units that exchange hands does not change much, so supply and demand are two sides of the same coin. When one goes up, it quickly depletes the other one, which then goes down. However, in the long term, supply is elastic, and construction eventually delivers new units to satisfy new demand. Unfortunately history shows that such delivery is often late, and results in hypersupply and eventually a recession.

So next time someone tells you that things will pick up in spring, or you see an article citing a modest increase in sales and claims that things have turned around, remember, one swallow does not make a summer.

Excel with the numbers and plots:

Leave a Reply