# An arm and a leg. What determines the price of real estate?

When I was a kid I used to collect soccer player cards every world cup. I remember that there were cards that were very hard to get by, and in school we would all fight for one of them. If a classmate had one, we would offer special favours and even part of our lunch for the card, and of course, we would also give many less rare cards in the ‘package’ offer. Those rare cards had low supply and high demand, and hence had a high price.

Prices of real estate are determined in pretty much the same manner. Supply and demand dictate prices. When the supply is low, and the demand is high, prices go up. When the supply is high, and the demand is low, prices go down. I find it incredible that when I discuss prices of houses with people, I realize that many of them don’t believe this. So let’s look at some stats to prove that this is true for Real Estate as well.

The goal is to compare price changes to supply/demand changes and see if they correlate.

## Price changes

As we discussed in a previous post, the House Price Index (HPI) is a rolling average that is a lot smoother than just average values. To compute changes in prices I looked at differences from month to month in HPI and noticed that the changes were still nosiy. This made sense, as monthly changes tend to be small and seasonality also plays a role. I then looked at whether the data got smoother by taking price changes over 3, 6 or 12 months. As the graphs below show, the price changes do get smoother, at the cost of a lag:

I then decided to use only the 12-month Y/Y HPI price change for the analysis below.

## Sales to Active Listing Ratio -> A measure of supply and demand

In Real Estate the demand is reflected by the number of sales. The supply is reflected by the total inventory of the product. Hence, if you take the ratio of sales to active listings (SAL), you get a very good measure of the supply/demand condition of the properties.

The REBGV defines the market as buyer’s market when the SAL falls below 12%, it defines it as balanced when it falls between 12-20%, and as seller’s market it is above 20% (Ref). Let’s see if that is true.

## SAL vs Price

I downloaded the Sales to Active Listing Ratio from Steve Saretsky’s website (https://stevesaretsky.com/vancouver-real-estate-market-stats/) back in October. The numbers appear to be only for the city of Vancouver (not everything the REBGV covers), but should be still very useful. Now let’s compare the House Price Index vs SAL for each unit type:

At first look, the trends make sense, when the SAL (blue bars) are high, the prices go up, when they are low, the prices go down. Let’s illustrate this in a different way, let’s look at the Y/Y price change (comparing Jan of one year to the Jan of the year before) vs SAL:

The trends are remarkably good, but you can see right away that the orange curve (the Y/Y change) is shifted to the right (as we saw above, the curve gets smoother but also shifts as we take Y/Y changes vs monthly changes). If you plot the correlations, they are not so good:

By shifting the Y/Y by different months, the lagging is removed:

Now the correlations get remarkably better:

Also of interest is to look at the intercept of the regressions with the Y-axis. This value basically should reveal the threshold for a ‘balanced market’. SAL values above that number would mean that prices will go up a few months later, and SAL values below that number would means that prices will go down a few months later. The intercept values for each unit type are:

 Unit type Intercept Detached 12.3% Townhome 13.7% Condo 15.8%

This means that what the REBGV says: “Generally, analysts say that downward pressure on home prices occurs when the ratio dips below the 12 per cent mark for a sustained period, while home prices often experience upward pressure when it surpasses 20 per cent over several months.” is not so far from the truth, although the threshold for condos is a bit higher, so SAL values below 15.8% should be considered buyer’s market. The SAL values from the last few months for each unit type are listed below, they should provide an idea of upcoming Y/Y changes in the HPI index (although remember that the regression has more points with high SAL values, so the error is high, so take it with a huge grain of salt!)

## Summary

I think we have demonstrated here that prices in Vancouver Real Estate do reflect the basic rule of supply and demand. We have also established that a small delay takes place, so that sellers adapt their prices based on the SAL of a few months ago. Interestingly, condos tend to have the shorter delay. My hypothesis is that price discovery is faster for condos, so prices adjust faster.

Now that we have established this, I leave you with a quote from Shiller which shows that supply and demand only reflect the psychology of the market. As this psychology changes (and it appears to be changing in Vancouver now), so will supply and demand:

The housing market levels we saw at the 2006 peak [in the US] were not, as so many imagined, the outcome only of fundamental forces affecting the rational demand for and supply of housing. Of course, home prices are set by the forces of supply and demand, as homebuyers so often say. The prices have to clear the market. But the factors influencing supply and demand include a lot of social and emotional factors, notably attention to the price increases themselves, a public impression that the experts know they will continue, and a predisposition to believe that they will continue. These factors will change with our changing culture. Understanding how social forces cause speculative market moves has been the major theme of this book. It is so difficult for most of us to figure out which moves are caused by sensible good reasons and expert opinions and which are caused by human imagination and social psychology. I hope that the argument to this point has made it clear that, as these major markets go, it is often largely the latter that drives prices.

In both the stock market and the housing market, people have only the fuzziest idea what these investments are really worth, what their prices ought to be. They may be able to judge whether one stock is overpriced relative to another, or whether one house is overpriced relative to another, but they just do not know how to judge the overall level of prices. Much more salient in their minds is the rate of increase of the prices, something that they talk and hear about a lot in a time of rapid price change, and that has subtle effects on their demand for speculative assets. As the price increase during a bubble goes on through time, people constantly reassess their opinions. People who thought there was a bubble, and that prices were too high, find themselves questioning their own earlier judgments, and start to wonder whether fundamentals are indeed driving the price increase. Many people seem to think that if the price increase goes on for years after some experts have called the price increase a bubble, then maybe the experts were wrong. And they then feel that there is no alternative to thinking that it is really fundamentals that are driving the increase, and that these fundamentals will go on forever. These are the phases that individuals who watch these markets go throughâ€” different individuals at different times.

[…]

When prices stop increasing for a long while, there is a gradually increasing discontent with this view. This gradually increasing discontent may cause stagnant or declining markets even when fundamentals are increasing. Markets may disappoint for stretches of years or even decades, as the post-bubble dynamics gradually play themselves out.

Shiller, Robert J.. Irrational Exuberance: Revised and Expanded Third Edition (p. 226-227). Princeton University Press. Kindle Edition.

Excel with the numbers can be found below. I downloaded all the statistics from Steve Saretsky’s website in October 2018. Since then I manually updated the numbers for Oct-Feb2019. Saretsky changed his website in January 2019, and some older numbers changed slightly, but did not seem to impact the overall conclusions, so I left them as they are.

Edit: I first posted this without explaining why I used Y/Y changes and Steven pointed this out, so I added that section and realized that my discussion of the lags being different wrt average vs HPI lags was erroneous and it really came from the fact that Y/Y price changes brings a lag in itself. I thank Steve for his comment.