« The "Cripes!" growth model | Main | A preliminary estimate for Canadian 2009Q2 GDP growth »

Comments

Feed You can follow this conversation by subscribing to the comment feed for this post.

Thanks for the graphs Stephen

just a suggestion though. Your price indice by city is set to percent change from peak. That's an interesting way to normalize the data but a little different from what most people present. It could lead to anomalies when different cities have peaks at different times, most noticeably in Calgary's prices. I think chart might look better if one uses percent change from a fixed date, although that method can also lead to anomalies.

Kind of a weird sample of cities there. The GTA makes up 42% of the sample, they ignore a huge area of the country in the middle (but not the Atlantic)...and Ottawa doesn't even include Hull! (that's just weird)

I suppose it's representative and doesn't make much difference, but since they obviously have data for Quebec, Winnipeg & Edmonton it seems odd to leave them out.

Nice to see Montreal housing prices holding up, since I'm a home owner here and about about to increase my mortgage to pay for renovations.

I remember reading an article about 6 months ago arguing that Montreal real estate prices were overvalued, but Toronto were not, despite the fact that Toronto prices are so much higher, because the rate of increase of real estate prices has historically been faster in Toronto than in Montreal. This seemed reasonable, until I realized they used as their historical period 1975 to 2000 (or something like that, I don't remember the exact years, or the author). Anyway, the period they chose to compare the long-term trend of growth in housing prices between the two cities corresponded to the period when so many head offices were moving from Montreal to Toronto. It seemed to me like a rather unfair comparason. I believe that from now on we'll see prices tend to rise slightly faster in Montreal than Toronto.

DR: If you go to the Teranet site, and look at the graph for Halifax, below the graph for the price index you can see the "sales pair count". It seems to be about 150 for May, and was much lower during the winter months.

I think the main reason they don't include more cities is that the sample size is just too small. Halifax is barely big enough, but they have to include it to get the Atlantics into the picture.

If you read (OK, skim) their methodology section, you also see they need to do a lot of procedures to try to get the index, all of which means throwing away some suspect data, and using up degrees of freedom.

The comments to this entry are closed.

Search this site

  • Google

    WWW
    worthwhile.typepad.com
Blog powered by Typepad