Last year the CMHC started releasing their quarterly housing assessments and I had some fun ridiculing their conclusions. Once again they’ve swapped out the analyst in charge of Victoria, which makes it the 4th analyst in just over year, and none of them are from here.
The new report for Spring 2017 is out, and by new report for Spring 2017 I mean released spring 2017 and actually only covering data up to Dec 2016. Again, how a measure is supposed to give an “early indication of potentially problematic housing market conditions” when the measure is at minimum 4 months out of date I have no idea.
They analyze the housing market based on evidence of Overheating, Price Acceleration, Overvaluation, and Overbuilding. Let’s go through their analysis in detail.
Their definition of overheating is a sales to new listings ratio of 80% or higher for the quarter. This is relatively sensible, although the threshold is quite high. Looking back at recent history, the hot market recently and the mid 2000s would have been identified as overheating. Note that the overheating evidence does not go higher than “moderate evidence” because it is only based on one metric (sales/list ratio > 80%).
CMHC defines price acceleration as follows. They take the 3 year moving average MLS Price, then they create a band of plus or minus one standard deviation. If the current MLS price is within that band, everything is OK. If it is above that band, there is evidence of price acceleration.
Seems on the surface of it to be logical. After all we know that prices have accelerated and sure enough it indicates price acceleration recently. Problem is that it is actually nonsense, and this method doesn’t identify price acceleration at all, it merely identifies price increases. This is clear when we look at the complete history of Victoria and mark the areas where the CMHC method identifies price acceleration.
Prices were not accelerating in most of the highlighted areas, they were simply increasing at a large enough rate to escape the 3 year average. So while the identification of price acceleration is useful, this method does not do it.
So when the CMHC says “evidence of price acceleration” what they are actually saying is “prices are increasing at above about 5% per year” which has nothing to do with acceleration. An example of acceleration would be if prices increase 5% in the first quarter, then 7% in the second, then 9% in the third.
CMHC rightly concludes we don’t have an issue with overbuilding here. Unabsorbed new build housing inventory is at record lows. That said, construction is on a tear lately and the number of units hitting the market will definitely be increasing soon.
This is where it gets interesting. First of all let’s remind ourselves that as recently as 6 months ago, the CMHC thought there was no evidence of overvaluation in Victoria. All was peachy and hunky dory. Then prices increased by some 5% and suddenly there is strong evidence of overvaluation? Clearly whatever method came up with that result is flawed.
The CMHC says that recent gains have resulted in “average prices strongly exceeding those supported by local fundamentals”. How did they come to this conclusion? Well they give this figure as Exhibit A where they compare income growth since 2009 to the average price.
Since price growth has outpaced income growth, their conclusion is that the market is overvalued. There are so many things wrong with that I don’t know where to start, but let’s go back to the whole point of this housing market assessment. The thing is supposed to alert people of problematic conditions with the assumption that problematic conditions could lead to price declines and a destabilized market. So what does comparing price gains with income gains tell us about the market? Pretty much nothing. House prices have outpaced incomes for decades, this is not a recent thing. By this measure the Victoria market would have been overvalued for as far back as we have data, and that makes the measure kind of pointless.
Of course, their method may be more complex than what they show – and likely is since they reference multiple metrics which they don’t talk about – but then why give such terrible examples to show that the market is overvalued? It’s not like there’s a shortage of affordability measures out there. RBC publishes a good measure of affordability regularly that shows Victoria detached houses at near record poor affordability levels and I’ve done the odd chart myself.
So is Victoria overvalued? I’d say in detached homes we are bumping into the limit of what locals can afford and that is good evidence of overvaluation. We’ve seen a massive drop in detached home sales last month and while part of that is due to low inventory, I believe part of it is also due to price exhaustion from local buyers.
I don’t disagree with CMHC’s overall assessment per se, but I don’t have any confidence in their method given what we’ve seen here in the last year. Firstly the data needs to be kept more up to date, and secondly we need to see some evidence that their method can identify problematic conditions when applied to the past. Would this method have identified previous peaks as problematic? Without some backtesting we have no idea if these housing market assessments are worth the bits they’re made of.