Mortgage rates are set nationally and not locally. Do you believe there is zero correlation between mortgage rates and house prices?
Furthermore, economic prosperity and conditions tends to be constrained much more by national borders than local borders. Do you believe the correlation between two local markets in the same country (Toronto vs Regina) is no greater than two local markets in different countries (Toronto vs Las Vegas?)
The argument that "national averages mean nothing" is usually followed up with the statement "it's different here".
To the first question, of course there is a correlation, but it is by no means uniform, which is what a national average of housing prices tries to convey. The health of a national market can be completely contradicted by the health of a local market, even though they share the same mortgage rate. If the correlation were direct then instead of only 11 areas of Toronto showing price increases (or 15%), the whole of Toronto would have (and none would have had significant price reductions as many areas of Toronto did), as well as Regina, Halifax, St. John, etc.
To the second question, I think that the gap is narrowing substantially and while I wouldn't go so far as to say that yes Toronto/Regina are no different than Toronto/Chicago, the very existence of have/have not provincial designation/transfers in our country shows that national boundaries are not necessarily indicators of economic health/interdependence - geographic might figure more prominently. Newfoundland has not been sharing in this recession to anywhere near the same extent as rural Ontario has.
It's foolish to use regional averages as representative of individual neighbourhoods. However, it is just as foolish to look at individual neighbourhoods while ignoring everything around them.
You use a comparative sentence, but don't follow through on the comparison ie: using neighbourhood averages as representative of regions...which of course would be equally foolhardy, and not something I'm suggesting at all. Real estate doesn't occur in a bubble and of course these are all factors.
What I'm saying is i can see no possible benefit, to anyone, in using a national average to convey health/sickness/prosperity/poverty when it comes to real estate. It's like taking East End Vancouver adding it to Forest Hill dividing by two and saying, hmmmmmm... seems like it's ok. It's a ridiculous and pointless exercise that gives us absolutely no information that is of any practical use or help.