MARCH 07, 2012
RBC has released their latest housing affordability report. It shows an improvement in affordability, which was quickly trumpeted by media outlets without a fleeting though of perhaps checking the methodology of the index to make sure it made sense. That's obviously way too much to ask of the media in Canada. Here are some gems:
Home ownership becoming more affordable- The Globe and Mail
Issues with RBC's affordability measure:
I've previously discussed in detail why I'd be very careful pinning beliefs of a stable market to a metric like this. In that post I outlined 7 reasons why the metric is flawed. I've also previously discussed why I have a huge issue with the assertion that since affordability has remained relatively stable according to these metrics, it means we have nothing to be concerned about. I think that's ridiculous. I won't rehash them here except to point out the major issue which is that the index assumes a stable 25% down payment over time. If people were putting 25% down 30 years ago and if people were putting 25% down today, yes their principal and interest costs would represent a comparable percentage of pre-tax income.
The problem, of course, is that the average down payment has shrunk over time, a fact discussed in the Summer 2011 Bank of Canada Review:
"Allen (2010–11) shows that from 1999 to 2004 most households with insured mortgages borrowed up to, or near, the maximum LTV ratio available at the time they purchased a home."
And since down payment requirements and savings rates have fallen over time, you can be sure that a stable down payment assumption is absurd.
Compounding the issue is the fact that since the 80s (incidently during the same time frame that RBC measures), interest rates have fallen fairly steadily. All things being equal, affordability should be well below historic norms. That they are at "normal" levels with interest rates sitting st rock bottom, in and of itself, should be enough to concern us.
The bottom line is not that the metric is useless, rather that like other measures of house price sustainability, this one has flaws. Decide for yourself if house prices in an area are sustainable at current levels by looking at a range of metrics, including price/income ratios, price/rent ratios, and price/GDP, supply/demand dynamics, demographics, etc (all of which can be found for different areas of Canada on this site by looking through the Primer articles). Hanging a bullish argument on an affordability metric alone would be foolish.
Bank of Canada affordability index:
I was asked to comment on this chart from the latest Bank of Canada review:
Once again we see that affordability trend is favorable and is actually well below its long-term trend. Unlike the RBC metric, this one actually assumes a stable 5% down payment. Yet the results don't pass the "common sense" test, so I did some digging.
Here's the methodology from the Bank of Canada website:
In order to better capture the value of both new and existing homes, (the house price portion of the index) is an equally weighted average of the Royal LePage Resale Housing Price (RLPHP) and the New Housing Price Index (NHPI)
So the change in house prices used in this index is weighted so that 50% of the change is derived from change of resale house prices through CREA (which is what Royal Lepage reports) and 50% is derived from the change in the New House Price Index (NHPI). I have a MUCH bigger problem with this methodology than I do with RBC's methodology. Here's why.
1) NHPI is quality-adjusted
As I have discussed previously, the NHPI is adjusted for changes in the quality of new homes. What this means is that changes in quality of the new homes in a year (or even just the size of the homes) are adjusted to create a constant comparison over time. Put simply, and as an extreme example, should all new homes built next year be twice as large and all contain the most modern amenities, despite the fact that the dollar cost of these homes would be much higher than in previous years, the NHPI would adjust for the changes in the size and amenities of these homes to show a much smaller gain in price (if any gain at all). In other words, in the mind of the statisticians who create the index, yes, they paid more for the homes, but it was better quality than homes built in the past, so the rise in price is not true inflation but represents an improvement in the quality of the homes, and therefore this should be accounted for.
I'm not debating the validity of the index as a component of the CPI; I understand the need to create a pure price change absent of changes in quality for measures of inflation, but to use this measure in an affordability index is, in my mind, a MASSIVE mistake. Put simply, the buyers of these huge luxury homes in the previous example have still paid much more in nominal dollars to buy these homes and it seems wrong not to account for that in a measure that purports to track average affordability of homes over time.
2) NHPI vs real-world measures of house price gains
To dig in a little deeper, below is a comparison of the growth in the NHPI vs the growth in the average resale house price in Canada (CREA) and the growth in the Teranet house price index, which also measures the growth in house prices but uses a paired-sale methodology. You'll note that the NHPI has significantly under-performed the other measures. This is largely because of the quality adjustments mentioned above:
I'll leave it to my readers to decide if it makes sense that the price of new homes built in Canada has significantly under performed the gains in the resale market. If not, how much faith should you put in an affordability measure that weights the increase in house prices 50% to this index?