A hypothetical example to show why housing stats are BS

Interesting point. I understand now why I never look at the averages. Instead, I save listings and analyze the changes I see.

REO point mathematically valid...mix matters...per sq foot compares would be better, but dont know if avail...this is one of problems Schiller attempts to solve in his methodology...what is interesting in the example is volumes doubled...higher volume is a sign of market strength (breadth and depth important as well)...whats happened in NY lately is volumes are quite depressed on y/y basis (sequential compares meaningless since the mkt is seasonal) suggesting clearing prices havent been reached...CA went from hi prices/hi volumes to high prices/drastically reduced volumes to lower prices/little volume to very reduced prices/more normal volumes from what i know...i dont believe we can conclude NY prices have stopped their precipitous decline until volumes more normal...i'm not close enough to the data to know what would be considered normal...perhaps the more data intensive/RE literate on the boards can fill in here

Well, in the last day there's been a ton of articles saying the Q2 "boom" in sales was still 50% off last years Q2 (I think).

REO - I think another way to prove your point is by considering how the number of closings in new condos (as a percentage of total sales) can skew average prices up.
Because the pre-development process in NYC is so much more expensive than any place in the known universe nearly all new development is "luxury." Accordingly, the price points of the new inventory is nearly always at price points above the median.
Accordingly, when new construction closings constitute 43% of all closings the median and average numbers will skew up. Now that new construction closings are less than 30% of all closings (and that number will continue to fall) it makes sense that median and average prices are falling. Even if the market were good right now the change in unit mix would have necessitated prices falling anyway. I realize that prices are falling because of economic troubles, but the problem is exacerbated by this change in unit mix.

So on the way up the value of an UES 2 bedroom built in 1985 didn't go up nearly as much as the sales data would have made one believe and now that prices are falling this same apartment is not going to lose value by as much as the numbers would have us believe (but it will still lose a significant amount of its value - just not as much as the stats will read).

This is the problem with Gov't housing statistics (OFHEO and its successors) and with NAR stats. Case-Shiller corrects for this problem.

Jazzman: without going point by point...... that was my point: not that things are higher or lower than the 'statistics' indicate, but that the statistics being presented (especially in the case of manhattan Coops and Condos, which is what most us us on this forum are really here to talk about), the numbers which we see thrown around aren't really worth much. So when people use these numbers to make their arguments, it's really GIGO.

REO - Sorry I wasn't more clear - I was agreeing with you that the numbers can lie - there are different reasons why they lie and I just wanted to focus a bit more on unit mix old to new rather than what you were focusing on unit mix studios vs 3 beds.

jm -

yes, I know, sorry *I* wasn't more clear. Just that my pint wasn't smaller vs larger ONLY. Or new vs old ONLY. or anything ONLY. But that in reality, we have ALL OF THESE THINGS and more, occurring simultaneously, and the concept that anyone has the ability to normalize the data is fiction.

One way of normalizing is by price per room (since coops don't have sq ft). It's not perfect but it's better than nothing. "In the blind man's world, the one eyed man is king."

And who's room count do you use? The broker's? Hell, even "official" room counts have changed over time to include or not include kitchens and other rooms. or worse, 1/2 rooms.