November numbers are out, and it’s pretty much as expected. The trend of an improving market continues unabated, with both MOI and Sales/List continuing along the lines they have been for years now. The release from the VREB is mostly as expected, expect for an unusal note by board president Guy Crozier.
“The good news for buyers is that pricing remains competitive and there are new listings coming onto the market every day.”
Now why would they mention competitive prices when the market is this active? Well it turns out the single family house median dropped a whopping 7.4% from October, or $43,500. Now this is most likely just a random occurrence, but fairly rare given a month to month swing of that magnitude has only happened 5 times in the last 20 years.
Also can anyone explain this graph to me? The VREB says it is sales to active listings, but calculating that value for November (573/2952) gives 19.4%, not the 25% that they come up with.
Just Jack uncovered a truly delicious tidbit in the comments on the last post which deserves more attention. As most of you know, the MLS Home Price Index was introduced a while ago to provide a more stable measure of house price changes than the median and averages, which tend to swing wildly month to month (November being exhibit A). The MLS HPI joins the much longer standing Teranet House Price Index. Both are based on sales pairs, so that the change in value of the same home can be compared, and thus both are dependent on having enough sales pairs to come up with a stable index value. In the past we’ve noticed already that the Teranet HPI starts behaving a little erratically once sales pairs in a given month drop below about 150.
Now here is where it gets interesting. The key difference between the MLS HPI and the Teranet HPI is that you can break down the MLS HPI into house types, and areas. So for example, one might want to look at single family homes in Gordon Head separate from all other houses and see how they performed. At first glance this is a sensible option to have, given how different areas can have totally different markets. The issue of course is that the more you divide your sales, the less data you have to come up with your benchmark home price.
Case in point: The VREB claims that the “benchmark” townhouse in Oak Bay increased in value by $23,300 (4.5% from October to November 2015). Sounds great right? That’s an annualized rate of 53%! Well Just Jack uncovered the truth behind that number.
There were no Oak Bay townhome sales in November. When I say no – I mean there were zero townhome sales. In October there was 1.
In other words, the increase in Oak Bay townhomes as stated in the MLS HPI is 100% pure fiction. There is no possible way you can calculate a benchmark price based on no data whatsoever. And by extension almost all of the subcategories by area are total fiction and should be ignored. I’m no statistician so I can’t say for sure how many sales pairs you need to make any statistically valid statements as to benchmark prices, but we can be sure that number is more than zero, and probably closer to 50. It’s pretty dishonest for the VREB to publish benchmark house prices that have no basis in fact.
So in case you find yourself reading the VREB news release, feel free to ignore pages 5 through 10, as they aren’t worth the pixels they’re printed on.