CMHC has been chiming in on Canadian housing market risks. Its Housing Market Assessment (HMA) framework rates risks for overheating, acceleration in house prices, overvaluation and overbuilding. Unfortunately, in my view, the ratings may be based on shaky foundations. Understanding why requires an appreciation of the HMA’s technical and statistical foundations. The methodology is heavy on general information but light on specifics. The lack of transparency on the specifics is what makes it a matter of faith as to exactly how CMHC arrives at its risk ratings.
Based on what little information CMHC has shared about the HMA framework, let’s consider the broad brush strokes about how it assesses risks. For the sake of brevity, let’s focus on how CMHC arrives at its risk rating for overvaluation (if you have questions about my concerns regarding the HMA’s other risk ratings then feel free to ask me about them in the Comments section below).
The HMA uses models to assess the risk of overvaluation. What’s important about these models is this: the worse they are at underestimating actual average prices in any particular market, the more likely it will be rated as “showing strong evidence of problematic conditions” (whatever that is supposed to mean). To my mind, this speaks more to the possible shortcomings of CMHC’s models in their ability to explain reality than it does “overvaluation.”
Personal after-tax income per person is used in one of the models to rate the risk of overvaluation. The idea underlying this model is that higher home prices can be supported by higher incomes and if prices grow much faster than incomes, then the housing market attracts a risk rating of “strong evidence of problematic conditions.”
The problem is that the growth dynamics for income and average price are fundamentally different. The differences in their growth dynamics are part of a normal housing market cycle, during which there are times when incomes grow faster than average prices and vice versa.
Moreover, average price is a terrible measure with which to gauge changes in the market value for homes. When average price grows faster than income, it may be because larger and/or more luxurious homes make up a greater share of total home sales rather than reflect any increase in market value and risk of overvaluation.
Additionally, the ratings are potentially stale-dated by the time CMHC publishes them. That’s because they’re based on the four most recent quarters included in the HMA rating. CMHC’s HMA labelled “third quarter” and published in July 2016 was based on data up to the first quarter of 2016. That means CMHC’s ratings were potentially based on modeling results that relate to the spring and summer of 2015. As a result, CMHC’s ratings lack relevancy from the standpoint of current market conditions.
The possible shortcomings in CMHC’s models, the use of average price data in its models, the lack of specificity and transparency regarding the basis for its ratings, and the staleness of data upon which the ratings are based: these are at the very root of my concerns about CMHC’s HMA risk ratings for overheating. You should share my concerns.